CONVERSATIONS WITH LATINOS:
What Makes Content Stick? How U.S. Latinos Navigate Today’s Digital Information Ecosystem

Scroll down to read the highlights and key takeaways in Spanish /Desplázate hacia abajo para leer los aspectos destacados y conclusiones clave en español.

[Press Release + Online Event Registration: Click here to join us for the launch of our new report on how U.S. Latinos engage with online content]

Download the Report PDF:

Introduction and Context

Social media and messaging apps have become deeply ingrained in our daily lives. From the moment we open our eyes in the morning, we are perpetually on our phones, tablets, or computers. Yet even as the average American checks their mobile devices upwards of 160 times per day, relatively little is known about what makes content stick while we are connected, what resonates with which audiences, and what many people are truly craving online – both consciously and subconsciously. 

Understanding the psychological, social, and media consumption drivers of engagement with information online is at the heart of the Digital Democracy Institute of the Americas’ (DDIA) mission to strengthen a healthier Internet for democracy. Building off of three foundational polls and a guidebook of best practices for creating prebunking content for Latinos, in 2025 DDIA set out to more deeply listen to Latinos about their experiences online, on social media, and in messaging apps.

In partnership with research firm Castillo & Associates and Latino fact-checking organization Factchequeado, with support from DDIA Advisory Council member and HIT Strategies co-founder Roshni Nedungadi, over the course of July and August 2025 DDIA conducted “Conversations with Latinos,” a series of qual boards and interviews geared at better understanding the following:

1. How U.S. Latinos consume content on social media, and what it is “about” content that captures their attention.
   

2. Who U.S. Latinos trust most on social media, and how that does or does not connect with what they choose to engage with for entertainment and to stay informed. Assessing this included assessing the characteristics of trustworthy accounts (green flags), and what makes people skeptical about certain accounts (red flags).
   

3. If/how participants recognize, interpret, and verify misleading or false information when exposed to it.
   

4. What Latinos perceive to be the biggest challenges to democracy and society.
   

5. How Latinos perceive AI’s use in creating or influencing information, its risks and benefits, and what their views are on AI-generated content.
   

6. Latinos’ perceptions of big tech’s role in democracy, platform regulations, censorship, algorithms, and online safety, with a focus on how people experience digital environments.

The findings outlined in this report can be used to shape the following:

• Information relevant for Latino influencers and creators committed to creating healthy, fact-based content online.
  

• Interventions and capacity-building strategies to enhance content production, credibility, and cultural competency.
  

• Training programs for journalists, content creators, scientists, policymakers and other front-line communicators working to reach Latinos.
  

• Policies related to social media, trust and safety, and Internet governance.

“Conversations with Latinos” comprised a qual board (an asynchronous qualitative online research tool similar to an online focus group) of 28 participants followed by in-depth interviews with 14 of the participants. 

The sample included a near equal mix of English-dominant and Spanish-dominant Latinos, women and men, from various age groups, and political affiliations. The group comprised Latinos living in the states of Arizona, California, Colorado, Florida, Illinois, New York, North Carolina, Nevada, and Texas. (A full breakdown of the methodology and participant demographics are included at the end of the document.) 

This study was designed to complement existing DDIA survey data, as well as other studies in the field, by adding nuance and revealing key “aha” moments behind U.S. Latinos’ digital behaviors. 

As a qualitative study, the findings outlined are not meant to be representative of all Latinos in the United States, but rather to illustrate insights and motivations that may be difficult to capture through usual quantitative research.

Key Takeaways Per Section

SECTION 1: Digital Habits, Content Consumption, and Go-To Sources

The Routine

Social media is so deeply ingrained in the daily routines of U.S. Latinos and society at large that content on platforms is frequently the first thing people engage with when they wake up, and often the last thing they look at before going to sleep. 

Most Latinos we spoke with shared that before setting foot out of bed, their routine included: 

• shutting off phone alarms, 

• checking the weather on weather apps, and 

• responding to messages in text or social media (with Spanish-speakers mentioning WhatsApp as well)

With little difference between age groups, gender, or political affiliation, in qual boards and interviews, Instagram emerged as one of the most consistently preferred platforms for routine use, valued for a variety of formats (photos, reels, stories, lives) and visual-first delivery. By many of the Latinos we spoke with, it was seen as offering information that was to-the-point while also being entertaining and informative, and it was perceived as being more curated for news than TikTok.

TikTok was mentioned nearly as often as Instagram as a go-to platform daily, especially by younger Latinos between the ages of 18 and 35. The app was appreciated for its immediacy and up-to-date content, especially on topics of interest or close geographic proximity to the participants (wellness, fitness, local events). While it was seen by many as “more laid-back,” credibility was a concern due to the high volume of user-generated content and AI-produced material.

Facebook remains important for community connection, especially among older users above the age of 40. Many cited liking Facebook for news, marketplace use, and for fostering and connecting with personal networks of friends, family, and community members. By the participants who referenced frequent use of Facebook, it was described as easy to navigate and reliable for local updates. Facebook groups, in particular, were seen by a few participants as desirable simply because of the opt-in nature of groups, which they found contributed to making their experiences more trustworthy and secure.

YouTube was valued for the depth of content, longer explanatory videos, and detailed breakdowns offered by producers. Participants also noted using YouTube as a background platform for music or multitasking. Some mentioned perceiving YouTube to be more trustworthy and well moderated (see more in next section).

ChatGPT, though a tool rather than a social medium, was often mentioned as “a fact-checking” and “ask anything” tool. While most participants still referenced using Google for proactively searching for content, AI search is clearly on the rise, and it was perceived as faster and more precise than Google or social media searches (see more in Section 3).

Araceli R., English-dominant, 36, Democrat, Florida: “Instagram. I've been using it for many years and follow everything that I want from friends, to business accounts, to news.”

Oscar G., Spanish-dominant, 21, Democrat, Texas: “Más que todo Instagram y TikTok son las dos principales que uso. Por lo general, TikTok es más diversión para mí, más para perder el tiempo viendo videos. En cambio, Instagram es más dónde puedo recibir contenido, agarrar información, noticias, y también me sirve como red social para hablar con otras personas, con amigos, y ver qué están haciendo en sus historias. En Instagram siento que el contenido es más de las personas, encuentro cosas más relacionadas con lo que me gusta, un poco más de la vida real. En TikTok, en cambio, me salen más videos que se sienten un poco falsos.”

Genesis I, English-dominant, 25, Democrat, New York: “I feel like it's [TikTok] very interactive and very up to date… people post in the moment. Even I do it.  If I see something, I’ll just post it on TikTok. It’s more like ‘in the now.’”

Elsie, Spanish-dominant, 65, Republican, New York: “La razón principal por la que tengo Facebook es para poder estar en contacto con personas que conozco, familia y amistades. Pero además de eso, lo uso para enterarme de lo que está pasando en el mundo.”

Participants highlighted a diverse set of accounts they routinely engage with online, spanning fitness, food, spirituality, lifestyle, and news. Personal routines intersected with participants’ broader social, political, and health-related content in their daily digital habits.

Accounts mentioned include:

• Fitness: JPG, Trevor Lane

• Cooking: Guga (chef)

• Health and Spirituality: Milagros Medium Terapéutico, a priest (name not remembered), two doctors (one neurosurgeon) who share medical developments

• News and Analysis: Citizen journalism on Instagram (local news posted by regular people), Charlie Kirk (debate clips on YouTube; noted as following debates more than the person)

• Individuals: Carlos Eduardo Espina, Fabricio Romano, Remy Parson, Shane Gillis

• Podcasts: Suburb Talks

• Lifestyle: Nice Kicks (shoe account)

What Makes Content Stick

One of the biggest questions of our time is what makes content stick in people’s minds while they’re navigating an overwhelming, often-polluted Internet. As a core part of this study, we dug extensively into what led people to stop scrolling and engage with content. 

Participants in qual boards and interviews were asked to scroll through their phones, and either submitted or directly showed us what led them to stop scrolling. As a part of this endeavor, participants were asked to describe which aspects of the content drew their attention and why. 

Main Point: When asked what content stuck out to them, most participants referenced the “topic” of posts first. When further prompted, many referenced content framing and the behavior of messengers. Notably, what sticks and is valued when people are passively scrolling seems different from what sticks and is valued when people are actively searching for something.

IN PASSIVE SCROLLING

Topics

Main Point: In passive scrolling, Latinos we spoke with valued content that touches themes or topics they are personally interested in - this comprised mostly a mixture of fun, light-hearted content and informational content. Entertaining content and informational content (here defined as content that taught people something new) naturally co-existed in many people’s feeds. 

When asked about what they hoped to come across on their feeds, Latinos we spoke with noted wanting to see things like local content and updates, entertaining content that makes them feel happy and ALSO informed, humorous content, content that will help them grow or teach them something new, and content related to things they enjoy, find relevant, or interesting (see more on the last point below).

Many noted they would rather not see content perceived as vulgar, obscene, sensationalist, explicit, violent, offensive, hateful, or focused on tragic or negative news. Several respondents expressed avoidance of political content. A number of participants mentioned avoiding advertising, gossip, and computer-generated or artificial videos. 

What drew attention because it was seen as “relevant” and “interesting” to the people we spoke with seems deeply connected to personal interests, hobbies, or culture.

• For many of the men we spoke with, regardless of age or political affiliation, topics of content that drew their attention included sports, fitness, content that taught them something or helped them become better people, cooking, cars, local crime updates, and immigration-related content.
  

• For many of the women, topics that drew attention included concerts and music, the health of kids and teens, cooking, local news, updates from friends or family, and immigration-related content.

Notably, immigration was brought up often as a topic of interest (one of the few not related to entertainment or hobbies), with participants across nearly all ages, political affiliations, and genders expressing immigration content often led them to stop scrolling. Multiple participants noted they’ve been watching content about the raids in the United States often. 

Spanish-dominant participants more than once noted they also engage with content “from their home countries.”

Salomon R., English-dominant, 39, Republican, Arizona: “I would say it really depends on my mood [the content he engages with]. For example, on the weekends, if I’m golfing, then I’ll be looking at golf content. Or if I’m watching sports, it’s going to be sports or fighting. During elections, it was a lot of politics. [He was] trying to see what was going on (around the riots and immigration issues).”

Abraham B., English-dominant, 28, Independent, Illinois: “I like cooking videos, watching police camera footage of incidents that take place, and… what else? Honestly, that’s about it. Oh, and workout videos, stuff like that.”

Gaby L., Spanish-dominant, 50, Republican, Georgia: “Quisiera más noticias que sean más positivas y que inspiren.” 

Ana E., Spanish-dominant, 57, Democrat, Colorado: “Pues a mí me gusta de todo un poco, pero en realidad ahorita estoy muy pendiente de lo que está pasando políticamente hablando de temas de inmigración, pero también me gustan temas de religión y de música. Entonces me gusta de todo un poco.”

Juan R., English-dominant, 45, Independent, Florida: “I avoid at all cost politics or sad news.” 

Leidy C., Spanish-dominant, 25, Democrat, California: “Evito noticias y cosas tristes o violentas.” 

Jesús M., Spanish-dominant, 38, Independent, Nevada: “Trato de evitar cosas políticas o los que tienen la voz computarizada de siempre.”

Framing of Information

Main Point: People are craving positivity and authenticity on social media. Joyful and/or informative and/or community-centered content stood out to participants as both engaging and desirable. 

Participants seemed most hungry for joyful, feel-good content on social media, as well as for informational, community-centered content. Latinos we spoke with noted that they will most often stop scrolling for posts that make them feel happy. They will also stop scrolling for urgent updates, especially local updates. 

People also valued informational content that teaches them something new, though often connected to topics related to their personal life. Latinos we spoke with, for example, cited engaging with content that taught them about plants they should have at home, or credit cards they should invest in, or data that sparked their curiosity.

Many Latinos expressly noted valuing well-organized, straightforward content that flows. Some Latinos we spoke with noted that “content can’t only give information and then loop back to certain points multiple times.” Clear communication seems important. 

People Featured in Content

Main Point: Friends, celebrities, or influencers as the subjects of content often catch attention. Participants we spoke with indicated they stop scrolling for content that features people they recognize.

Many of the participants mentioned that they will stop scrolling when the subject of a post is someone they recognize, be it a sports figure, an actor or actress, or a family or friend. Names of famous figures cited included soccer player Cristiano Ronaldo and Venezuelan actress María Gabriela de Faría.

Messengers 

Main Point: During passive scrolling, many Latinos express valuing engagement with individual over organizational accounts. Behaviors and characteristics of individual accounts appreciated included evidence-sharing, authenticity, proximity, and approachability.

• Individual Accounts Over Organizational Accounts:
  

  • Like in the prebunking study DDIA conducted in 2024, Latinos we spoke with often expressed a preference for engaging with content produced by individuals rather than by organizations when passively scrolling their feeds. Many people feel more connected to individuals, especially when they share similar backgrounds or experiences. Individuals were many times seen as more approachable and less “saturated” than organizations.
    

  • Still, almost everyone we spoke with could not remember the names of specific content creators or influencers they liked or appreciated. Often, people remembered the behavior of the creator or influencer, like whether the person showed evidence to back up their opinions or had a specific aesthetic (see more below), more than the name or profile of the person. Many Latinos also noted they value “authentic” or “genuine” voices.
    

  • Behaviors or components of individual accounts that were widely valued by participants included::
    

    • Evidence-sharing: What the producer of the content showed them: for example, people who expressed opinions backed by specific citations, facts, or evidence that they pinpoint or visibly shared as part of the post.
      

    • Proximity of the Messenger: Both geographic and contextual/cultural proximity are important. This included physical closeness (e.g., a pastor they like hearing from, a local journalist in their community, or a friend or family member), but also shared context or identity. Respondents described they follow voices who were also immigrants or descendants of immigrants, who liked the same sports, or who had lived through similar struggles they themselves are facing.
      

    • The Messenger’s Aesthetic: Latinos we spoke with mentioned an aesthetic that makes you think people have their lives together is notable to them. For example, a few participants noted that messengers can’t have a cluttered background, that presentation is important, that very low quality videos project less credibility.

Ilse M., English-dominant, 37, Republican, Texas: “If it's a person, it's more their demeanor, or how they're speaking or presenting, presenting the video in and of itself.”

Yanina A., Spanish-dominant, 26, Republican, Texas: “Información detallada, videos, pruebas de lo que están hablando, porque nunca he visto un noticiero que diga o que hoy un muchacho murió y tal, pero lo hablen y no pase nada. No, siempre hay una reportera en el lugar de la escena del crimen, o hay una reportera hablando con un profesional. Y eso es lo que me gusta, pues que te dan la información detallada con pruebas.”

Formatting

Main Point: During passive scrolling, people valued individual accounts that present information in a “clean,” straightforward and well-produced way. They tended toward human-centric videos or imagery and, again, valued the showing of evidence to back up assertions.

Though formatting of videos was not a primary component many Latinos we spoke with noticed first or noted as being important to them consciously, people illustrated some preferences for specific production qualities or formatting types, including for:

• Short, straightforward videos: Many noted liking video reels, but none cited long-form content, or written articles. There were notable preferences for faster and easier to consume content: quick, direct images and videos with concrete information. Ongoing research in the field is actively trying to better understand the role of long-form content in media consumption diets.
  

• Well-produced, human-centered content, over commercial content: High-quality production was also mentioned several times: good sound and quality resolution of videos and images seemed important for viewers. Many of the Latinos we spoke with cited disliking AI-generated voices. Many noted they valued real human voices for authenticity.
  

• The showing of evidence through imagery or citations: Many were drawn to content that shows evidence: supporting videos or images. The people we spoke with appreciated when messengers showed proof to back up their opinions or analyses. Headlines seemed to be a key hook, especially for news-related content. Many indicated preferring headlines paired with attractive, relevant images rather than plain text.

Edenilson E., Spanish-dominant, 26, Independent, California: “Lo que me hace quedarme, por ejemplo, es como que alguien haga un video hablando de noticias o de cualquier cosa, y que muestre attaches como pruebas. Esto me hace quedarme.”

Angel A., English-dominant, 31, Democrat, Illinois: “As far as the first account that I showed you, one uses a lot of bold lettering or bold text on the image superposed on the image. And it's just kind of very straightforward direct information to you, and that's easy to read so I'll show you what it says: ‘Bad Bunny wears unreleased Adidas slides on late night,’ and you know, you just swipe through it, and it's just straightforward. You know exactly what you want to see.”

Oscar G., Spanish-dominant, 21, Democrat, Texas: “A mí personalmente no me gusta cuando los videos se enredan mucho, cuando hablan demasiado al principio y quieren abordar el tema despacito. Prefiero que digan: vamos a hablar de esto y va a pasar esto… También creo que influye mucho la voz o la persona que lo está diciendo. Eso me hace seguir viendo, porque hay voces que transmiten bien la información y logran atraparte. En cambio, cuando la voz es muy aburrida, o solo pasan muchas imágenes sin sentido, eso es lo que me haría dejar de ver o adelantar el video.”

• Words that Latinos who participated in the Qual Boards responded with when asked what they associated with “information that entertains me:” (in no particular order)

AESTHETIC, COMEDY, CONNECTION, FUNNY, EYE-CATCHING, EVENTS, FUN, INFORMATIONAL, INTEREST, LAUGHTER

• Words that Latinos who participated in the Qual Boards responded with when asked what they associated with “information that captures my attention:” (in no particular order)

EASY, EXPRESSIONS, FLASHY, FUN, GENUINE, IMAGES, INFORMATIVE, INTERESTING, KNOWLEDGE, LEARNING, POSITIVE, RELATABLE, SHORT, SKILLS, UNDERSTANDABLE

IN ACTIVE SEARCHES

Though many people today inform themselves about things happening in the world via the passive consumption of social media content, it is important to note that people do search for information and in doing so value truth and credible content. 

Journalism is still seen as one of the best providers of that credible content in active searches – consumption of traditional media is still there, especially in cases where people are looking to proactively find trustworthy information, or when they would like to learn about what is going on in the world. 

It is important to note that no single outlet seems to be fully trusted; people construct a mosaic of information across platforms, individuals, and verification practices. Even with “trusted” accounts, participants emphasize the need to double-check, they acknowledge information gaps, and remain aware of media bias.

“Information-Seeking” Preferences

Main Point: Most participants described engaging in active or hybrid search behaviors when they want to get informed – the majority either combine scrolling with follow-up searches or search for information directly. Social media, especially Instagram and TikTok, are common entry points to informational content, and when people’s interests are piqued by specific information, they indicate engaging in active searches most often on Google, YouTube, or ChatGPT to verify information or get more context. 

• TikTok and Instagram were the most frequently referenced social media for staying informed. TikTok is valued for its immediacy (lives from the scene, and the feed surfacing relevant info), while Instagram is valued for visuals, reels, and the ability to show different perspectives. YouTube was trusted by some for greater “veracity,” longer explanations, and direct access to creators. Facebook and X, though mentioned less frequently for active search, were mainly valued for local information or professional news accounts.

Ilse M., English-dominant, 37, Republican, Texas: “I don't think I've ever gotten content that was directly from WhatsApp, but in regards to Instagram or TikTok, I give them the  same amount of credibility. I will give these two sources even more credibility than the news, because they are coming from "real people." 

Oscar C., Spanish-dominant, 32, Republican, Republican, North Carolina: “Uso mayormente para esto [to get informed] TikTok. Siento que esta red social se convirtió en mi “mano derecha” cuando quiero informarme de cualquier cosa con un video corto objetivo y sin necesidad de escuchar un discurso previo.”

Edenilson E., Spanish-dominant, 26, Independent, California: “[He prefers] TikTok porque deja hacer live a personas que andan donde se genera la noticia. En general para noticias aquí en California sigo a una persona que se llama Richard noticias.”

• Google and Search Engines: People often use Google and other search engines both to get informed and to fact-check. Other search engines mentioned were Yahoo, MSN, or Apple News to cross-check events and gather updates.
  

• ChatGPT: Used by several participants to fact-check, clarify political processes, or explain complex issues. It is valued for being straightforward and easy to use, but typically complements other sources rather than serving as the first point of contact.

Ilse M., English-dominant, 37, Republican, Texas: “If something catches my eye, I’ll go check it out on TikTok, Google, or ChatGPT.”

Angela G., English-dominant, 50, Republican, Florida: “I usually have the search bars or apps on my phone like Yahoo, Google. I usually navigate and read what articles and news are headlined.”

John P., Spanish-dominant, 34, Democrat, North Carolina: “Por lo general me gusta informarme por YouTube o directamente voy al navegador de Google.”

Trustworthiness of Sources

Main Point: Though most of the Latinos we spoke with noted they don’t trust any individual or organizational social media account 100%, citing elements such as bias, corruption, and self-interest, the element of trust remains important for people actively looking for information. Participants’ trust in sources during active search appears to be context-dependent, shifting based on the topic, the source, the type of content, and the behavior of the messengers. When it comes to platforms, YouTube emerged as the most trusted of all social media platforms by many of the participants. 

Conditional Trust: The degree to which many of the participants we spoke with trusted people or accounts varied based on the degree to which accounts shared evidence and whether content matched their personal beliefs.

Journalism: People may not be consuming a news-heavy diet, but when they are actively seeking out reliable information about current events, they still indicate going to well-known news outlets, both local and national, as well as .gov and .edu websites, local news outlets, and citizen journalists.

When discussing news sources they go to for finding information about politics and society, participants highlighted one Spanish-language U.S. outlet, Telemundo, which they valued for its on-the-ground coverage, documentation, and community presence. Several participants cited independent Spanish-language journalists, such as Carla Angola, Paola Ramos, Carlos Loret de Mola, Joaquín López-Dóriga, as examples of reporters who provide reliable and independent perspectives. These references suggest that Spanish-language news is often trusted when it feels close to communities or when independence from political agendas is perceived.

By contrast, participants mentioned a wider range of English-language outlets, including CNN, NBC, CBS, Fox, BBC, and local news stations, but tended to approach them with more skepticism. Many described not trusting any single outlet “100%,” instead preferring to cross-check information across multiple English-language sources. While these outlets remain central to how participants get informed, they are often filtered through doubt, fact-checking, or even supplemented with alternative channels such as Reddit communities or ChatGPT.

• Mainstream and International Media: Seen as direct, wide-reaching, and generally professional, though bias awareness is high: BBC, Fox, CNN, NBC, CBS, Telemundo, the CNN News App, the Fox News page on Instagram, and MSN.com were all mentioned.

• Local News Outlets: Valued for proximity and community relevance, according to participants. Outlets specifically mentioned included: KSL (local radio), Chicago/CBS, the Our Chicago Reddit feed for hyperlocal info.

• Citizen Journalists or Political Pundits: Chosen for consistency, on-site presence, and ability to simplify complex issues, according to participants. Specific names mentioned included: Carlos Eduardo Espina, Benjamin Zamora, Paola Ramos, Aaron Parnas, Carlos Loret de Mola (Mexico), German Loaiza (Salvadorian reporter), Yusmaria Mateus (Venezuelan nanny that works in the U.S. and teaches kids how to speak Spanish and Venezuelan customs), VPI Houston, Richard Noticias, Carla Angola, Hispatlanta. Others mentioned included Candace Owens, Charlie Kirk, Ben Shapiro, Jordan Peterson.

• Community and Alternative Sources: Reddit (especially local threads), newsletters, and neighborhood groups also surfaced as relevant, especially for local or niche issues.

YouTube: When asked which social medium they found most trustworthy in general, most participants indicated they found YouTube the most trustworthy, depending on the specific channels. 

Oscar G, Spanish-dominant, 21, Democrat, Texas: “Cuando me quiero informar reviso mi página de Apple News en mi teléfono. Sin embargo, muchas veces encuentro noticias o reportajes en Instagram y luego las busco en línea o me meto a la cuenta que lo publicó para saber más. Uso Instagram y X o Twitter ya que en mi opinión son las plataformas que tienen más cuentas profesionales de noticieros y encuentras por así decirlo menos desinformación. Sigo en especial a BBC world para saber más noticias en el propio instagram.”

Christian A., Spanish-dominant, 43, Democrat, Colorado: “Para informarme voy a páginas de noticieros. Prefiero Instagram. Sigo páginas como Telemundo, CNN, etc.”

Gaby L., Spanish-dominant, 50, Republican, Georgia: “Me informo a través de cuenta verificada, como @carlaangola, @hispaatlanta.”

Yanina A., Spanish-dominant, 26, Republican, Texas: “Sigo a Carlos Eduardo Espina, un muchacho que habla directamente de lo que pasa, y lo encuentro en TikTok. Siempre que deseo tener información de algún tema, yo entro a TikTok o ChatGPT.”

Main Point: Many Latinos we spoke with indicated relying on cues for determining trustworthiness and credibility of the information and accounts with which they chose to engage. These included the showing of evidence to back up opinions, verification badges, popularity and comments, opt-in features and a track record of past accuracy. Very few people were doing deep dives into accounts to ascertain their credibility. 

• Evidence and Proof: Latinos we spoke with placed high value on content that shows detailed information, videos, and proof, to back up opinions. This applies to both organizations and individuals.

Genesis I., English-dominant, 25, Democrat, New York: “If it shows enough facts, like video proof and text proof. Also, if they are someone that works in the field that they are talking about.”

PJ T., English-dominant, 27, Independent, Florida: “Studies backing up the info presented. Again, cited info showing the data presented is fact.”

Ilse M., English-dominant, 37, Republican, Texas: “A source does not necessarily need to come from an account with a verified symbol on it, but if it has one, then I do feel like it is more trustworthy.”

Salomon R., English-dominant, 39, Republican, Arizona: “Data that backs it up, I’ll then follow up on Google search. I take it with a grain of salt until there is confirmed news about it on TV or Google search. Normally I trust Fox news or something like that.”

• Verification Badges: Verification badges and check-marks remain a signal of credibility despite changes made by specific platforms to the parameters required to gain verification, which many people do not seem aware of. Many people we spoke with shared that they place more trust in verified accounts.

• Popularity: Follower engagement and comments are still being used as a means of discerning credibility and trustworthiness. Accounts with larger followings that have been “around for a long time” are often perceived to be “safer.”
  

• Intent to Inform: As we’ve seen time and again, Latinos are not gullible. Many of the people we spoke with valued the messengers’ clear intention to inform rather than sell or promote. They noted being able to tell when accounts were trying to push certain points of view.
  

• Opt-In Features: People seem to trust communities they can opt in to (Reddit, Facebook groups, WhatsApp groups)

Ana E., Spanish-dominant, 57, Democrat, Colorado: “When you are subscribed to something private, you have to go through a series of questions, they have to give you permission to be there, and that makes her feel more secure, more precaution, a gusto, barrier to entry, opt in.”

• Track Record: Past accuracy was important to many of the participants in this study. Multiple noted that they began to distrust certain influencers or accounts if they noticed that posts were inaccurate or wrong, even when the information was later corrected.

Edenilson E., Spanish-dominant, 26, Independent, California: “Por ejemplo, el influencer este que le comento, que se llama Carlos Eduardo Espina. Yo lo seguía al principio. Era como que tiene buen léxico. Tiene buena buena oratoria y yo decía, está todo bien, pero no me comenzó a gustar porque distorsionaba la información. Por ejemplo, quitaron una ley de ayuda a los niños de la frontera. Un ejemplo, al día siguiente ponía no ya no fue que la quitaron fue que entró en proceso de ver si la quitan. Entonces te está quitando. Si me entiendes como que a veces sí, entonces a veces no. Entonces te quedas con la incógnita.”

• Words that Latinos that participated in the Qual Boards responded with when asked what they associated with “information I can trust:” (in no particular order)

ARTICLES, FAMILY, GENUINE, NEWS, OFFICIAL, RESEARCH, RELEVANT, RELIABLE, RESPONSIBLE, SERIOUS, STUDIED, VERIFIED

Persuasive Elements of Content

Main Point: First-hand perspectives, such as direct interviews and on-the-ground footage increased the persuasive qualities of content. Contrasting viewpoints, including seeing different sides of a story, also prompted rethinking of perspectives.

When asked if they had seen content on social media that helped them change their minds about a specific issue or topic, most participants acknowledged that they had to some degree. Below are three examples that were shared:

• Israel-Palestine: The participant gained empathy and understanding for Palestinians by seeing direct interviews with people experiencing the conflict.
  

• Migration Issues: Seeing content about the poor treatment of immigrants challenged this Republican participant’s trust in authorities meant to protect society.
  

• Donald Trump: The participant’s perceptions of President Donald Trump shifted from negative to positive after exposure to TV and social media narratives highlighting perceived economic benefits of his administration.

Elsie, Spanish-dominant, 65, Republican, New York: “He cambiado un poco mi punto de vista con todo lo que está sucediendo alrededor de la inmigración. Hay tantas cosas que uno no sabe si son justas o injustas, porque faltan detalles o información. Y eso hace que uno pierda confianza en las autoridades, porque te preguntas si realmente están haciendo las cosas como deberían.”

Trust in Latino-generated vs. non-Latino-generated content is not uniform: it divides along lines of language, cultural connection, and perceived expertise of the messenger. 

• Latino content is valued for relatability, accessibility, and sometimes proximity to events, but critiqued for being more emotional, humorous, or less fact-checked.
  

• English/non-Latino content is often perceived as more authoritative or fact-based.

Elsie, Spanish-dominant, 65, Republican, New York:  “Creo que si tuviera que escoger, me inclinaría más por lo latino, porque quizá sea más fácil entender ciertas cosas en español que en inglés. Pero tampoco pensaría que lo que dice una persona norteamericana tenga menos valor; no creo que haya una gran diferencia. Todo depende.”

Edenilson E., Spanish-dominant, 26, Independent, California: “El contenido en inglés es más verídico. Siento que los latinos buscan más distorsionar la información. En cambio, el contenido en inglés viene de personas que saben un poco más de lo que están hablando.”

Representation

Many people could not pinpoint specific people who they felt represented them online. That said, some noted specific figures that reflect their personal preferences, issue interests, or political or ideological leanings. These included:

• Activists and Advocates: Carlos Eduardo Espina (trusted for defending immigrants and sharing a Latino/immigrant background), Aaron Parnas.

• Political Leaders: Nayib Bukele (President of El Salvador, valued for policies that benefited family).

• Thinkers: Jorge Bucay (psychologist, appreciated for his insights).

• Entertainers: Hermanos de Leche (comedy creators who speak frankly without claiming absolute truth).

• Religious Figures: A priest (name not remembered).

MAIN POINT: Many times, the Latinos we spoke with noted certain traits that drove their perception of whether a messenger represented them personally. These ranged from cultural or community connections, to the person’s defense of immigration or other topics important to them, to the communication style of the messenger.

• Cultural and Community Connection: Shared Latino or immigrant identity.

• Advocacy: Defense of immigrant rights and policies benefiting the community.

• Authenticity and Communication Style: Speaking directly, using simple words, and avoiding claiming “absolute truth.”

Approachability: Feels like the person is talking “to” them.

Genesis I., English-dominant, 25, Democrat, New York: “I think his name is Aaron Parnas. He's very big on keeping people updated on things that aren’t really talked about, like family situations and personal stories… We’re bombarded with so much information, but he explains it in a way that helps. He advocates for the community and tells the smaller, personalized stories that news outlets often miss.”

Spanish-dominant Latinos indicated being more likely to follow and engage with Latino creators, while English-dominant Latinos tended to have a more mixed feed. Still, across groups, Latino voices remained important for certain topics and types of content.

• Spanish-dominant Latinos consumed more content from Latino creators, generally motivated by language comfort, cultural connection, and ties to their country of origin.

• English-dominant Latinos showed a more mixed consumption, balancing Latino and non-Latino content.

Notably, even participants with predominantly non-Latino or mixed consumption said they prefer Latino voices specifically when it comes to immigration topics, as well as for comedic content.

Yanina A., Spanish-dominant, 26, Republican, Texas: “La verdad, sí, veo más contenido de latinos, como un 90%. Eso es porque creé mi TikTok y mi Instagram en Perú, entonces me sale contenido de esa región… Me gusta el contenido en español porque lo entiendo mejor. Hablo inglés, pero me siento más cómoda en español.”

Abraham B., English-dominant, 28, Independent, Illinois: “I’ve been seeing more non-Latino content. The only Latino content I come across is around immigration status or things related to ICE. That’s been more recent, before it wasn’t like that.”

Genesis I., English-dominant, 25, Democrat, New York:  “For political content, it’s mostly Latinos, I’d say the majority is Latino. But my general content is very mixed: Latino, Asian, and others… Right now, politics online feels very Latino-focused, especially around immigration.”

SECTION 2: Online Harms

The Biggest Threats to U.S. Democracy

Main Point: When asked to name the biggest threats to democracy today, almost all of the 28 qual board participants cited some version of online harms in the digital information environment, including misinformation, disinformation, and the role of social media in spreading false or misleading content that undermines trust, as well as the political use of power, a sense that leaders act in self-interest, overreach, or fuel polarization rather than represent the people. Other issues such as racial injustice, freedom of speech, and foreign policy were also mentioned.

Specifically, Latinos offered up the following words to describe what they perceived to be the biggest challenge:

• Misinformation

• Disinformation

• Social Media

• Political Divides

• False Media

• Political Parties That Make Changes To The Country Based On Beliefs That Alienate People

• Racism

• Discrimination Related To Immigration Rhetoric

• Radical Ideas And Laws

• Executive Orders

Abraham B., English-dominant, 28, Independent, Illinois: “People not caring enough to vote, misinformation, and, I would say, political divide.”

Araceli R., English-dominant, 36, Democrat, Florida: “Social media. We have so much information going around that isn’t accurate. And people tend to share information without looking for further information.”

Isle, M., English-dominant, 37, Republican, Texas: “The biggest threat to democracy today FROM MY PERSPECTIVE, is that representatives don't really vote like if they were speaking for the people that they represent. I feel like they vote for whatever suits them best personally, and their main job is to REPRESENT.”

Salomon, R., English-dominant, 39, Republican, Arizona: “Social media and how they make people look. Having opinions and convincing them.”

In follow-up interviews, we also sought to understand what people see as the biggest “challenges society is facing” today. Many participants noted the following:

• Immigration: Immigration was seen as a primary national challenge, especially for Latino and immigrant communities. The challenge was often associated with fear of deportation and government-led enforcement. A few participants mentioned the role of Republican stances and rhetoric in increasing a sense of exclusion.
  

• Misinformation and/or Fake News: Across demographics, we saw widespread concern about rumors, false news, and misinformation shaping public perception.
  

• Racism and Social Division: Racism, specifically anti-immigrant sentiment, was a recurring challenge brought up. Responses often tied this worry to political figures, including specific citations to President Donald Trump, and broader cultural attitudes.

• Erosion of Values: We heard general concerns about violence, lack of respect for life, and moral decline in society.

Abraham B., English-dominant, 28, Independent, Illinois: “I’d say one of the biggest things is social media. Everything gets thrown out there, and people don’t know what to believe. There’s also a lot of political stuff, which depends on opinion and perspective. And immigration is another big problem.”

The Role of Social Media in Unifying or Dividing Society

When asked whether online spaces allow for nuanced discussions or not, respondents consistently recognized that social media can be both unifying and divisive. 

• On one hand, many saw social media as enabling community-building around shared interests and offering ways to connect with friends and family.
  

• On the other hand, participants noted that these same spaces are highly divisive, marked by hateful content, polarizing debates, and limited opportunities for deeper, more meaningful connections. 

An important reflection from a participant was that social media has shifted from being a unifying space to one increasingly marked by distrust.

Although the sample for this qualitative study was small, some differences in how people interpreted these dynamics emerged. Democratic-leaning participants tended to emphasize the dual nature of social media, both its unifying and divisive qualities, while Republican-leaning participants were more likely to highlight the community-building and conversational opportunities, with less focus on division. These patterns should be interpreted with caution, but they nod to how political orientation may shape perceptions of division in online spaces.

Genesis I., English-dominant, 25, Democrat, New York: “Social media works both ways. It can build community by tailoring your feed to your interests, like food or fashion, and connecting you with others who share them. But it can also amplify negativity, feeding you hateful content that feels like a personal attack, especially when people can hide behind anonymity.”

Ana E., Spanish-dominant, 57, Democrat, Colorado: “Yo creo que antes las redes sociales se unían más, pero ahora, por todo lo que ha pasado, se están desuniendo. ¿Por qué? Por la desconfianza en una de las plataformas que más me gusta usar y donde tengo mis grupos: WhatsApp. En realidad, WhatsApp ya no es 100% confiable, porque ahí también te pueden robar información. Eso ha pasado. Aun así, me gusta porque me mantiene enlazada con mi familia a través de los grupos: ahí compartimos y recibimos información.”

Ilse M., English-dominant, 37, Republican, Texas: “I think online spaces can feel like a safe space. Sure, someone might attack you in a comment, but usually it’s just a matter of saying, ‘No, I believe this because of that.’ In fact, I think social media opens up conversations you might not be having with people in real life.”

Polarization and Fear of Online Backlash

Main Point: Participants frequently shared that they avoid expressing opinions online due to fear of hate, backlash, or confrontation. 

Concerns included protecting their personal information, preserving their mental health, and maintaining friendships. One participant explained that they felt more comfortable sharing opinions on X (Twitter) precisely because their account was anonymous. 

PJ T., English-dominant, 27, Independent, Florida: “On Twitter I feel more comfortable sharing my opinions because my account doesn’t show my face. I don’t follow anyone, I don’t have followers, it’s just for information and content. That anonymity makes me feel like I can speak freely without repercussions.”

• A small group expressed no issue with posting their views online, while a few participants emphasized that they prefer to share opinions in person, especially with trusted friends and family. 

Abraham B., English-dominant, 28, Independent, Illinois: “If I’m having a conversation in person, it’s different, I’m engaged and open about my opinions and beliefs. But online, I don’t think it’s worth it. Too many people will just disagree and end up hating on you.”

Topics That Trigger Taking Sides

The issues most often cited as pushing people to “take a side” were political debates, elections, and immigration. Other polarizing topics included feminism, Israel-Palestine, policing, health, and broader cultural debates.

Main Point: Politics and elections emerged as the most frequently cited polarizing themes. Participants described feeling that when it came to these topics, online spaces left little room for nuance or middle-ground positions.

Ana E., Spanish-dominant, 57, Democrat, Colorado: “Yo creo que en temas de política sí [tiene que escoger un lado.]. Sobre todo durante las campañas, cuando los candidatos todavía están en promoción y los partidos empiezan a sacar los trapitos al sol. Yo ya me inclinaba por cierto partido, pero a veces uno no se da cuenta de quién es realmente esa persona hasta que ves lo que el contrincante publica. Para eso sirven mucho las redes sociales, porque cuando eres del lado opuesto te muestran cosas que quizá no conocías. Eso me hizo investigar más y también actuar de otra manera. Esto me pasó en la candidatura para la presidencia de Estados Unidos.”

Salomon R., English-dominant, 39, Republican, Arizona: “I guess I’ll take it back to COVID. Back then, the whole debate around vaccines felt so polarized. For me, because of the uncertainty, I didn’t want my family, especially my kids, to get it [the vaccine.]. But for the sake of being able to travel and work, I had to do it. I got mine about six months before my wife, and when she asked if she should get it, I told her, ‘I don’t feel any different(...)’ Either you were taking the shot because the government told you to, or you believed the government was trying to kill you and refused it. There was nothing in between. Same with elections, no middle ground, no space for nuanced views.”

Genesis I., English-dominant, 25, Democrat, New York:“I felt it [having to take a side] a lot during this past election. To me, it came down to whether a party was really for the people or not. One side felt more about advocating for hate toward certain communities, and as a Hispanic immigrant myself, I couldn’t agree with that. I felt more connected to the other side, because of their focus on education, experience, and advocating for people. It made me feel like I had to take a side, not just stay quiet.”

• Some participants noted that certain events may pressure people to take sides, and that social media adds an element of shame for not doing so. One respondent reflected on the Israel–Palestine war:

Angel A., English-dominant, 31, Democrat, Illinois: “This whole Israel versus Palestine thing has made me feel like I have to pick a side. I’ve always been pro-Palestinian, but I just never really shared anything online anymore as much as I used to. Now, with people shaming those who stay silent, I’ve definitely felt like I need to show solidarity and share my thoughts.”

• Immigration and identity-based stereotypes was another theme where participants felt compelled to choose sides. Online debates, rumors, and stereotypes triggered strong responses.

Ilse M., English-dominant, 37, Republican, Texas: “Back when that whole Coca-Cola thing happened, people were saying, ‘Coca-Cola called ICE, so you can’t drink Coca-Cola anymore.’ I thought, well, I guess I can’t drink Coca-Cola anymore.”

PJ T., English-dominant, 27, Independent, Florida: “When people make statements with stereotypes, like ‘Hondurans are prone to being violent’ or ‘Puerto Ricans are such criminals,’ I know it’s not true. But then on the Latino side you also see things like ‘every white person is racist.’ And sometimes what they point to doesn’t even seem racist to me.”

Other Polarizing Issues

• Gender: Topics such as gender roles, abortion, and feminism often surfaced as moments when people felt compelled to take sides.

• Police: Expressing support or respect for law enforcement online was described as potentially provoking backlash or tension in relationships.

• Media: The way different media sources frame the same story through partisan lenses (e.g., Republican, Democrat, Independent) was seen as reinforcing polarization and making it harder to find neutral ground.

What Latinos Perceive to Be Harmful Online

Main Point: When asked to offer input on what content people perceive to be most harmful online, respondents frequently cited false or misleading content, hate speech, online bullying, and fear-mongering content. Immigration narratives and racist rhetoric were seen as common harmful content specifically targeting Latinos.

Specific answers included:

• “Things that are not factual”

• “Information that is not accurate”

• “Information that is misleading”

• “Videos that are made up and are not real”

• “Content that is misleading. For example AI that now produces voice and is easy to think are real.”

• “Content that purposefully disinforms or guides people in the wrong direction”

• “Content that can access your information without consent”

• “Hate speech to any community”

• “Bullying”

• “Content that is fear-based”

• “Hate speech and extremism. This fuels discrimination.”

Specific examples of real harmful content they saw online: 

• “Coca-Cola video claiming the company was denouncing and deporting its own workers”

• “Videos about immigrants destroying our country”

• “ICE raids”

• “A Spanish lady yelling at a peaceful protester to go back to their country”

• “Graphic videos”

• “Doxxing and swatting”

• “Immigrant kids being taken away from their parents”

Harmful Narratives Targeting Latinos

In the qual boards and interviews, participants were asked to share examples of harmful content they had seen targeting Latinos on social media. Narratives about immigration and racism emerged as examples almost unanimously.

• Immigration-Related Exclusion: Participants described narratives portraying immigrants as not belonging in the U.S., often attributed to “Americans” or White people asserting “this is our land.” These narratives framed immigrants as “taking jobs,” “not paying taxes,” and “not caring about the country.”

Angel A., English-dominant, 31, Democrat, Illinois: “Yeah, definitely. Absolutely. Narratives like, immigrants who come here don’t really care about the country. They take jobs.”

Salomon R., English-dominant, 39, Republican, Arizona: “Right now, with the whole immigration issue, you hear people saying immigrants are taking jobs, that they’re here not paying taxes, and that it’s bad for the country. I think a lot of people believe this without really doing research on what it actually means for Latinos to be here.”

Oscar G., Spanish-dominant, 21, Democrat, Texas: “Por lo general los americanos siento que son los que principalmente tienen esa necesidad de decir: ‘Yo soy de aquí, este es mi país, esta es mi tierra, y tú no perteneces aquí.’ Te hacen sentir menos.”

• Racist or Aggressive Content: Many pointed to racism and hostility appearing in comments, videos, conversations, and even news. Political leaders, including the U.S. president and his team, were cited as amplifying exclusionary narratives that then circulate in media and online.

Ana E., Spanish-dominant, 57, Democrat, Colorado: “Últimamente he visto mucha agresión y mucho racismo, y eso no me gusta (...) yo lo he estado observando generalmente en gente americana (...) lo he visto en redes sociales, en videos, en conversaciones, en noticias. Inclusive, a veces hasta en la información del mismo presidente, de su gabinete o de las personas que lo rodean.”

Abraham B., English-dominant, 28, Independent, Illinois: “I’ve seen racist comments online, especially around immigration. For example, when ICE raids were happening, I saw posts saying things like, ‘This is what should be happening, keep these people out of the country.’”

• Figures Driving Division: Participants also mentioned divisive public figures and influencers, even when they did not follow them directly.

Angel A., English-dominant, 31, Democrat, Illinois: “Race and immigration are especially hot topics that fuel conflict in the comments. A lot of this comes from anonymous users, but divisive figures also pop up on my feed, even if I don’t follow them. People like Ben Shapiro, Candace Owens, and others, people who, in my opinion, try to create division with their ideas.”

• Crime and Inferiority Narratives: Narratives about Venezuelan and Latino criminality were described as magnified online, while others pointed out how Latinos themselves sometimes internalize inferiority in comparison to Americans or Europeans, reinforcing stereotypes.
  

• Harmful Content Generated by Fellow Latinos: Some participants also reflected that Latinos may contribute to harmful narratives themselves, for example through vulgar music, exaggerated stories shared for views, or out of fear.

Perceptions of Disinformation Targeting Latinos

In interviews, when asked whether participants believed Latinos are specifically targeted with disinformation, 11 of 14 participants affirmed they believe Latinos are more targeted than other communities. They pointed to a variety of reasons why Latino communities may be engaging with disinformation:

• Immigrant Status and Language Barriers: Many noted that immigrants often arrive in the U.S. without full knowledge of how institutions work and may face language barriers, which makes them more likely to believe misleading information.
  

• Cultural Tendencies Toward Trust: Some described what they saw as a cultural pattern of accepting information without questioning it. As one participant put it, Latinos sometimes believe too much in everything they hear, which contributes to the rapid adoption and spread of false content.
  

• Exclusion and Targeting: A few participants framed disinformation as part of a broader effort to marginalize Latino communities, placing it alongside harmful narratives aimed at other groups, such as Black communities.

Edenilson E., Spanish-dominant, 26, Independent, California: “La mayoría de latinos acá son latinos que han venido al país a buscar sus oportunidades, su forma de avanzar. Quizás porque no tienen quizás más conocimientos, entonces quizás es más fácil engañarnos.”

Elsie, Spanish-dominant, 65, Republican, New York:  “Puede que sí [latinos reciben más desinformación], porque tendemos a ser bastante crédulos y a creernos todo lo que nos dicen. Por ejemplo, por la manera en que se criaron”

Araceli R., English-dominant, 36, Democrat, Florida:  “The motivation, I believe, is just to show that we don’t belong here, that’s all. But it’s okay, we’re not everybody’s cup of tea”

Oscar G., Spanish-dominant, 21, Democrat, Texas: “He tenido muchos amigos afroamericanos en la ciudad y siento que también a ellos les intentan poner mucha información falsa. No sé exactamente por qué, pero sí he visto amenazas, por ejemplo, de que en ciertos departamentos iban a empezar a entrar a las casas porque habían encontrado afroamericanos haciendo cosas.”

Perceived Spreaders and Motivations for the Spread of Disinformation

When asked about the sources and motivations behind disinformation, 10 participants offered a range of perspectives in interviews. 

MAIN POINT: The most frequently cited motivations for the spread of disinformation were: content creators seeking likes, money, or virality; political actors spreading fear or shaping narratives for electoral gain; and motivations tied to racism or domination over Latinos.

Foreign or external actors were also mentioned as potential spreaders, particularly in cases where harmful content seemed profit-driven or showed signs of advanced technological manipulation.

• Content Creators: Several participants pointed out that some Latinos themselves produce or amplify disinformation, often for likes, money, attention, or virality.

John P., Spanish-dominant, 34, Democrat, North Carolina:  “Mi ingenuidad me hace pensar que es gente que simplemente quiere hacer likes, o se quiere hacer dinero a punta de movimientos en línea. Pero puede ser que sea también de más allá… no lo sé.”

Jesús M., Spanish-dominant, 38, Independent, Nevada: “Viene de gente que simplemente quiere más likes o más atención. Y es muy, muy fácil, se va a escuchar feo, pero es muy fácil engañar a una gran población latina. Basta con poner una información escandalosa y va a llamar la atención, sea cierta o no. En muchas cuentas que he visto, para mí, simple y sencillamente es una batalla de likes o de ser famosos por un pequeño momento.”

• Political Actors: Participants perceived political parties and politicians as key drivers of divisive narratives. One explained that it is usually the opposition of the moment that spreads negative or misleading information, pointing out that when Trump was president, those against him promoted damaging narratives, and when Biden was in office, the same dynamic played out in reverse. Another participant emphasized that anti-Latino narratives often come from the Republican side, particularly around immigration. 

Ilse M., English-dominant, 37, Republican, Texas: “I don’t know, I think it might just be really strong political figures, very biased, very black-and-white in their views.”

Edenilson E., Spanish-dominant, 26, Independent, California: “Yo pienso que, en lo personal, la información falsa o engañosa viene de los partidos políticos, especialmente de la oposición del momento. Por ejemplo, cuando estaba Trump, los contrarios a él le tiraban lo malo; y cuando estaba Biden, pasó lo mismo al revés."

• Individuals with Harmful Intent: People who feel superior or act out of racism, seeking to make Latinos feel like they do not belong.

Oscar G., Spanish-dominant, 21, Democrat, Texas: “Creo que eso viene de los mismos grupos que siempre han estado en contra de ellos [Latinos], que los han considerado inferiores... En mi opinión, muchas veces es solo por hacer el mal, por esa satisfacción de hacer sentir mal a otros. Cuando son extorsiones, ahí sí hay un interés económico, pero en muchos casos creo que es simplemente por molestar o arruinar.”

Salomon R., English-dominant, 39, Republican, Arizona: “I think it started with someone targeting the community, and once it reaches Latinos, they’re used as the front-runners to influence or change people’s views. So I believe the source probably comes from non-Latinos.”

• For-Profit Accounts: Participants noted that one of the challenges with disinformation is that its origins are often unclear. Some believe it comes from other countries with advanced technological capabilities, where harmful content is created to generate money, whether through likes, attention, or scams.

Ana E., Spanish-dominant, 57, Democrat, Colorado: “Mucha gente viene de otros países. Son personas que tienen mucho conocimiento en tecnología y pueden manipular todo esto, por lo que es de cuidado. La mayoría lo hace por motivos económicos: quieren beneficiarse porque al ganar seguidores o vistas, las plataformas les pagan. Otros son scammers que obtienen tu información y te la roban… Otros simplemente lo hacen para difamar. Hay de todo.”

Exposure to Disinformation

Main Point: All participants affirmed that either they, or someone in their family or close circle of friends, have been exposed to disinformation at some point. Immigration-related disinformation, including false information about raids, restrictions, and enforcement emerged as most common today. Scams on messaging apps were also commonly cited.

The below examples were the most top of mind, or most easily recognized as disinformation, by participants. Various recurring themes emerged: 

Migration and Border Issues

Participants most frequently described having been exposed to disinformation about immigration, ICE, and border enforcement, including false information about raids, restrictions, and threats that created fear and uncertainty in their communities:

• Viral videos of ICE raids

• Claims that immigrants’ bank accounts would be seized

• Rumors about border closures or misinformation about what officers were asking for at checkpoints

• Stories that residents could not return to the U.S., or that students’ parents might be detained when picking them up at school

• The Coca-Cola case, claiming the company collaborated with ICE to target undocumented immigrants

• Claims portraying all immigrants as criminals, regardless of their legal status or personal circumstances

• Sensationalist or fabricated stories, such as people at Alcatraz being “fed with dogs” or rumors about riots in Los Angeles that never actually happened

Edenilson E., Spanish-dominant, 26, Independent, California: “Decían que los agentes de inmigración estaban en cierto lugar aquí en California, en Downtown Los Ángeles, en las calles Wilshire y Serrano. Yo llamé a mi mamá, porque ella trabaja por ahí en un hotel, y le dije: ‘Mamá, inmigración está ahí.’ Me preguntó quién me lo había dicho, y le respondí que lo vi en TikTok, en un video de Carlos Espina. Normalmente él dice cosas reales, así que pensé que era cierto. Pero pasó por ahí y no había nada. Aun así, mucha gente lo creyó desde un principio.”

Jesús M., Spanish-dominant, 38, Independent, Nevada: “Recuerdo hace unos meses cuando empezaron a decir que Coca-Cola había llamado a migración para que fueran por sus empleados aquí en Estados Unidos. Pero nunca decían en qué planta, ni entrevistaban a las personas. No sé quién lo empezó, pero los videos estaban muy orientados a que la gente le tomara cierto odio o venganza a Coca-Cola.”

Genesis I., English-dominant, 25, Democrat, New York:  “One example was about the criminalization of how people migrated to the U.S. Before, you were considered a criminal only if you had been convicted of certain crimes. But then the narrative shifted to say that anyone who entered the country ‘illegally’ was a criminal. That change made it seem like people who risked their lives to migrate, seeking safety or a better future, were automatically criminals, putting everyone on the same level.”

Scams and Fraud

Another frequently cited topic of disinformation was scams, especially on WhatsApp and other messaging platforms. People cited:

• Fake job offers targeting immigrants

• WhatsApp scams promising services or financial help

• Hacking

• Fraudulent alerts such as fake USPS messages

Ana E., Spanish-dominant, 57, Democrat, Colorado: “Hay muchos engaños por todos lados. Lo primero que se me viene a la mente son propuestas de trabajo falsas. Durante mucho tiempo vi anuncios que decían: ‘No necesitas estudios, trabaja desde tu casa y gana 1,000 o 2,000 dólares a la semana.’ Esa información es ridícula. Ahí te das cuenta de que no es real. Generalmente te ponen un link para que entres, des tu información… y todo eso es mentira.”

Health Misinformation

Some participants recalled health-related misinformation, such as claims that vaccines cause autism or distorted reports about political leaders’ statements on COVID-19 (e.g., Trump’s alleged comment about drinking bleach).

Elsie, Spanish-dominant, 65, Republican, New York: “He visto desinformación sobre las vacunas, como que podían causar autismo o que no eran seguras. Yo no creo que sea cierto: tengo dos hijos, solo uno es autista, y si las vacunas causan autismo todos los niños vacunados lo tendrían.”

Politically Charged Content

Finally, participants noted politically motivated disinformation, including manipulated or out-of-context statements attributed to public figures and stories designed to discredit political opponents.

Participants pointed to specific platforms where they encounter harmful content. Facebook and TikTok were mentioned most often as spaces where disinformation spreads widely, while WhatsApp was primarily associated with scams and fraudulent messages.

John P., Spanish-dominant, 34, Democrat, North Carolina: “Ahorita, con los temas de migración, ha habido mucha información falsa. Mi mamá es inmigrante (...)  Muchas veces ella recibe información en Facebook, no sé por qué, pero es ahí donde circula tanta información falsa, y me la reenvía. Yo la veo en estado de alerta, y soy quien la calma diciéndole: ‘Esto puede tener algo de verdad, pero esto es totalmente falso.’ Sí, más que todo, el problema es Facebook.”

Fact-Checking

Participants described clear routines for verifying information online where they had doubts. While Google was the most common starting point, participants also drew on a variety of other sources, from ChatGPT to news outlets and personal networks.

MAIN POINT: People indicated they will share fact-checks mostly when the information personally affects them or the people around them. Most people cited they would share verifications with friends and family directly as opposed to publishing them on social media. Some people, when they suspect fake news, will report it directly to the platforms, but many don’t understand that to be an option.

Primary Fact-Checking Methods

• Google Search: The most common first step, often to see if claims appeared in trusted reporting. Participants noted using Google to “cross-check multiple sources,” compare coverage, and even run reverse image searches. 

Oscar G., Spanish-dominant, 21, Democrat, Texas: “Casi siempre voy a Google y busco directamente la información. Veo 3, 4, 5 pestañas con la misma información y pienso que tal vez sea verdad, pero nunca me quedo 100% seguro.”

Jesús M., Spanish-dominant, 38, Independent, Nevada: “Cuando me llega información o videos en español sobre política, casi te puedo decir que siete de cada diez son refutables. Los reviso rápido en Google y, si no aparecen en la página oficial correspondiente, sé que no han sucedido.”

Abraham B., English-dominant, 28, Independent, Illinois: “Oh, you could always reverse the image back on Google. And image reverse search. And you can see where the original picture came from, and then how they took a photo from something and made it.”

• ChatGPT: Used for quick answers, summaries, or an initial check. Participants claimed to pair ChatGPT with Google or official sites for confirmation.

John P., Spanish-dominant, 34, Democrat, North Carolina: “Hace poco escuché que Trump había hecho un ‘mega plan’ o algo así. Lo único que hice fue entrar a ChatGPT y pedirle un resumen de qué se trataba esa ley, lo más concreto posible. Quería ir directamente a la información clara de lo que había visto.”

• Official Websites: Government portals and specialized sites.

• Media Outlets: CNN, Fox News, NBC, WSBN/Channel 7, and local stations are consulted. Several participants emphasized comparing across outlets to account for bias.

Araceli R., English-dominant, 36, Democrat, Florida: “If it’s something related to the government and I don’t see it on Gov, CNN, or Fox News, then I don’t believe it’s really happening. Even when it is reported, I compare because media outlets can have their own bias. At the same time, social media is often faster, I’ll sometimes see a story on social media 15 or 30 minutes before CNN or Apple News report it.”

• Friends & Family Cross-Check: Asking trusted people, especially those with direct knowledge of a situation 

Genesis I., English-dominant, 25, Democrat, New York: “When news first broke about Ukraine, I asked my Ukrainian friend if her family was okay. I do the same with my Palestinian and Israeli friends. Asking directly gives me peace of mind, because then I know what’s really happening.”

• Social Media Verification (YouTube, TikTok, Comments): Some participants used social platforms themselves to verify information. This included following trusted creators on YouTube or TikTok (valued for their “trajectory” or credibility), scanning posts to see if claims were circulating more broadly, and even reading the comments, where other users sometimes corrected or contextualized false information.

Ilse M., English-dominant, 37, Republican, Texas: “I’ll go through the comments, like, okay, what are people saying about it? A lot of the time people in the comments will say, ‘That’s false’ or ‘That’s not true, what’s true is this.’ And if I scroll through and still don’t know what’s going on, then I’ll probably just go ask ChatGPT if it’s true or not (...) Yeah, that’s my go-to source. [referring to ChatGPT]”

Together, these practices show how Latinos actively triangulate information across platforms, media, and personal networks, yet many remain skeptical, rarely placing full trust in any single source.

SECTION 3: Artificial Intelligence 

As part of this project, we also set out to hear about Latinos’ experiences with artificial intelligence (AI) in real-world settings online.

Perceptions About AI and AI-Generated Content

Main Point: Many respondents have mixed feelings about AI, even as quite a few noted AI has made their day to day lives a bit easier. While many see AI tools themselves as neutral, they note AI can be a positive or negative based on who uses it and for what purpose. Some participants also expressed uncertainty about AI’s presence in their lives, recognizing that it is “probably right under [their] nose” in ways they may not fully notice or understand.

In line with previous research done by DDIA, when asked if they perceive AI-generated content to be positive, negative, or neutral, nearly all of the people we spoke with indicated thinking it was neutral. Most understand AI-generated itself is not either positive or negative, but rather can be used for good or bad purposes. Some respondents noted whether AI-generated content is good or bad is heavily dependent on the creator of the content.

Participants noted the biggest benefits of AI being:

• Greater and quicker access to information

• Increases in productivity

• Efficiency in personal and professional matters

• Faster decision-making

• Time saved

PJ T., English-dominant, 27, Independent, Florida: “It [AI] just makes life easier. Instead of me having to waste my time and go to Google and then look through links. It kind of just gives me its opinion based on Google. And then if I'm reading through and I don't really like it, I'll ask a specific question. So it's like having a conversation with you, except it's like with a robot that pulls stuff from Google.

John P., Spanish-dominant, 34, Democrat, North Carolina: “Yo pienso que la inteligencia artificial va a ser más beneficiosa que negativa (...) En lo económico y en temas legales puede ser muy útil. Yo estuve involucrado en una situación legal y le pedí a ChatGPT que me ayudara; me lo resolvió rápido, algo que un abogado me hubiera cobrado caro y habría tomado mucho más tiempo.”

Angel A., English-dominant, 31, Democrat, Illinois: “Yeah, I think it’s made certain processes easier. And it’s probably right under my nose. I don’t even know that it’s AI working for me.”

Concerns included:

• AI being misused or exploited for harm

• AI being used to create false, hateful, or manipulative material

• Fear that AI can “do anything” or that the tools’ creators can gain excessive political and institutional power

• Worries about AI replacing jobs, “making people lazy,” or “removing the human touch” from day to day life

• Apprehension about AI gathering personal data without consent

Oscar G., Spanish-dominant, 21, Democrat, Texas: “Siento que mi principal preocupación es que tiene acceso a tanta información y tanta capacidad de pensar que también está en un límite de usarse mal, y que pueda echar a perder redes sociales, información, la sociedad, todo.”

Jesús M., Spanish-dominant, 38, Independent, Nevada: “Una de las mayores preocupaciones es que los propietarios [de estas tecnologías] lleguen a tener más poder y control, incluso dentro de lo democrático o legislativo. Porque si ellos tienen el poder o la facultad de tomar decisiones, nada les impediría alterar la información que recibimos al hacer una búsqueda para que todo se incline a su favor.”

Ilse M., English-dominant, 37, Republican, Texas: “My biggest concern with AI is misinformation. Not necessarily from AI itself, because I don’t think AI is actively plotting against us, but from how people use it. I’ve seen examples, like fake weather forecasts where AI makes it look like there’s an earthquake. Somebody could believe that.”

Main Point: How people perceive AI-generated content when they come across it online or on social media aligned with their expectations of how they went about engaging with information and in which spaces, especially between passive scrolling and active search. When passively scrolling, respondents indicated feeling “annoyed” coming across low-quality AI-generated content, whereas when actively searching, many trust AI to give them trustworthy, unbiased content.

When more passively scrolling through social media for entertainment or engagement, many respondents indicated feeling annoyed or put off by the prevalence of what has become known as AI slop, usually referring to low-quality, mass-produced, often generic content created using gen-AI tools. Most did not appreciate seeing AI-generated imagery or videos, as they perceived those to be inauthentic and artificial. In passive scrolling, AI-generated content was also generally not seen as trustworthy. Many participants indicated a strong preference for seeing stories generated by real people. They noted finding human-generated content far more credible than AI-generated imagery of video. 

Yanina A., Spanish-dominant, 26, Republican, Texas: “La inteligencia artificial la usan para mentir y la usan para decir verdad, entonces no confío al 100% (...) Cuando lo habla un ser humano, es cuando más me da confianza en un video.”

Genesis I, English-dominant, 25, Democrat, New York:  “It [AI generated content] makes me question things a little more. I feel like I always have to double-check and ask myself, ‘Is this real? Is this actually happening?’ AI content can look and sound very real, but it still gives me doubt. So I go to news outlets or articles to confirm whether it actually happened.”

Salomon R., English-dominant, 39, Republican, Arizona: “No, I don’t trust it [AI-generated content] as much, because we don’t really know who’s behind it.

When actively using AI for search, which seemed common among many of the Latinos we spoke with, however, participants indicated they tended to trust AI-generated responses. Respondents were more frequently using AI to search, aggregate, and fact-check information, and as they perceived to have control over those searches, many shared they found ChatGPT-generated responses credible. 

Adoption of AI

ChatGPT remained the most-cited AI tool by participants, in line with findings from the two DDIA polls conducted in 2024, wherein participants indicated the same. Other platforms mentioned included Grok.

Younger users between the ages of 18 and 35 have become active adopters of ChatGPT, whereas older users above the age of 50 indicated some trepidation with the tool, expressing fear of exposure and loss of control over information. 

Most common uses cited included:

• Fact-checking (quick verification of online information).

• Academic Support (studying, simplifying research, solving problems).

• Business Growth (branding, graphics, captions, videos, social media management).

• Emotional Support (therapy, learning assistance).

Siri and Alexa appear as secondary, more-task-specific assistants.

Jesús M., Spanish-dominant, 38, Independent, Nevada: “Cuando quiero a veces corroborar, sí he llegado a preguntar una que otra vez a ChatGPT: ‘Oye, acerca de esto, ¿qué encuentras en tu base de datos?’ Y me aparecen ahí las opciones.”

Araceli R., English-dominant, 36, Democrat, Florida:  “Oh yeah, I use ChatGPT a lot. I use it to make flyers and promotions. I also use it for images. For example, I work in music and entertainment, so I’ve made cartoon-style designs, like a penguin version of myself for a logo, or a birthday graphic for a friend. I also use it sometimes for documentation, like resumes. Even Microsoft Word now lets you type in what you need, like a lease agreement with names, amounts, and terms, and it generates the whole thing for you. That’s crazy to me. If everything in artificial intelligence were to go down, I don’t know how we’d go back to the old ways.”

Main Point: Conversations with the participants surfaced that people think they are coming across AI-generated content far more often than in the past couple of years. 

When asked how much content generated by AI they think they come across on social media every day, many people shared that they see “a lot” or “more and more nowadays.” Respondents’ answers to “how much” ranged from 5% of content they see being AI-generated, to 60% being AI-generated. Some noted they see at least 2 - 3 posts per day that are AI-generated. 

These incredibly varied responses to the question of how much AI-generated content people see on a daily basis can suggest both uncertainty about what is AI-generated and variation in people’s experiences and feeds. It also underscores that perceptions of AI prevalence may not align with actual levels of exposure.

Perceived Ability to Detect AI-Generated Content

Main Point: Generally, the participants we spoke with think they know when content they see is AI-generated. They use mostly visual and voice markers to assess whether content is AI-generated.  

• Regarding imagery or visual markers: flags included imagery being “too perfect,” “too unrealistic,” “too 3D,” “almost always looking the same,” or looking “a little off.” People expressed that they were able to recognize unrealistic textures, cartoon-like appearances, or unnatural backgrounds relatively easily.
  

• Regarding the person depicted in the content: flags included the messenger’s “voice being too slow,” “lips not syncing to audio,” or “voices sounding too perfect” or “identical across videos.” Multiple people noted they are able to recognize distinctive or repetitive AI-generated voices, Some also mentioned spotting mistakes like people having more than five fingers.
  

• When presented with three videos — one of which featured an AI-generated voice — some interviewed did not notice the AI voice and even considered the tone of the voice “more professional.” as if it was produced by an organization (in the video’s case USPS) rather than a person. Out of the 28 people interviewed, only 6 said they preferred the AI-voice video for sharing. Among those who did recognize the use of AI, several mentioned that the robotic tone made the voice feel less trustworthy. 

PJ T., English-dominant, 27, Independent, Florida: “As of right now, I can tell the difference because AI isn’t that phenomenal. It still feels robotic, eerie, not human. But it’s close enough to human that I think some groups of people will be susceptible to getting caught off guard and believing AI content is real content.”

Oscar G., Spanish-dominant, 21, Democrat, Texas: “Siento que la inteligencia artificial intenta asimilar mucho lo que nosotros [humanos] hacemos, o sea, el movimiento, la voz, todo eso, pero desde la voz hasta las texturas, sobre todo en los videos, se nota. Siento que la textura es la que más me dice esto no es de verdad, o sea, ninguna cámara toma ese tipo de fotos, como que todo tiene una similitud a una caricatura.”

Main Point: Though many of the people we spoke with thought they could detect AI when they saw it, many participants did not think others, especially older adults or children, can tell AI-generated content from “real” content. 

There was a recognition that while “some people can detect AI-generated content, others don’t care or don’t notice.” 

When asked how people felt when they realized content they were seeing was AI-generated, people’s responses ranged from saying they felt normal or are “used to it,” to that it “made them feel weird,” and “like nothing is real anymore,” to feeling frustrated or deceived by AI-generated content.

Edenilson E., Spanish-dominant, 26, Independent, California: “Para mí hay mucha gente que consume inteligencia artificial y piensa que es real, pero no es real.

[Por ejemplo] Mi mamá me dijo: ‘Uy, hijo, vi una noticia de que una muchacha en Guatemala tuvo 5 hijos, 5 hijos de verdad.’ Incluso hasta guardó el video para enseñármelo. Cuando regresó de trabajar me dijo: ‘Vení a ver.’ Y yo le dije: ‘Esto es inteligencia artificial, lo están usando para poner cosas que no son, y no es real.’

Ella reaccionó como cuando alguien te miente: se sintió frustrada, como diciendo ‘creí’.”

Participants also reported not being able to tell what was AI-generated content themselves in the past, but that they are more aware and discerning now. 

Various participants noted they were surprised by how realistic AI-generated content has become, with some noting the AI made content less trustworthy. 

AI Labeling

Main Point: Across almost all 14 interviews, when asked if AI-generated content should be labeled, a majority of respondents indicated they thought it should. Participants noted that platforms should mark AI-generated content to “prevent confusion.”

Ana E., Spanish-dominant, 57, Democrat, Colorado: “Pues a lo mejor una alerta sí [sobre si las plataformas deben indicar cuando es un contenido generado por IA]. Porque de hecho hay muchas personas que desconocen que existe [IA]... No creo que la mayoría de las personas se puedan dar cuenta de la diferencia entre un contenido generado por inteligencia artificial y uno que no.”

Ilse M., English-dominant, 37, Republican, Texas: “Yes, I think that’s a responsibility [to label AI content] that should be done both by content creators and by social media, that there should be a warning saying that it’s [the content] AI.”

SECTION 4: Big Tech and Regulation

Who Decides What We See on Social Media

Main Point: Most participants recognize platform algorithms as a primary deciding factor for what people see on their feeds, and several note that platforms intentionally push content that drives high engagement, which many recognize can drive the spread of harmful content. 

Main Point: The belief that “phones are listening to us” is repeated and recognized as feeding related ads or content.

Several participants we spoke with say content is now dominated by ads and those who pay. Multiple participants acknowledge their own searches and clicks influence content.

Jesús M., Spanish-dominant, 38, Independent, Nevada: “Para mí, se ha escuchado mucho eso de que las aplicaciones tienen su algoritmo de repeticiones, tiempos, etcétera, y que en base a eso te mandan las cuentas para mantenerte entretenido. No sé si sea 100% real, pero sí tiene mucho sentido que sea así.”

Araceli R., English-dominant, 36, Democrat, Florida: “The algorithm has this kind of magic. Like, we could be talking right now with my phone next to me, and maybe it picks up bits and pieces. Later on, when I’m on social media, I’ll start getting little things (...) It feels like whenever you’re speaking about something, you’ll always come across related content. For example, if I’m talking about health, then health-related posts will start showing up.”

Thoughts on Broader Online Risks

Main Point: When asked which are the biggest perceived online risks, participants noted everything from misinformation and the promotion of violence online, to privacy and surveillance, to hacks and scams, to targeted harassment. 

Many participants cited fears of being tracked online or having their personal information or data stolen. 

Main Point: Respondents generally perceived themselves to be well able to detect harmful content online. Groups many participants perceived to be most vulnerable to online harms included children and older adults, with some people sharing they think everyone online is vulnerable to risks at some time.

How to Solve for a Harmful Online World

When asked how online risks could be solved, participants’ answers revealed a mix of expectations, skepticism, and personal responsibility. 

Main Point: Latinos we spoke with mostly placed the responsibility for securing a healthy online world on social media platforms and tech companies, suggesting they could create stronger filters, ban harmful accounts, enforce copyright, and introduce stricter review processes before content is shown. Several participants emphasized that individuals and parents also play a crucial role in keeping the Internet secure. A few noted the importance of users being “real” online and not simply chasing likes, framing part of the solution as a matter of personal responsibility.

Some pointed to filters, verification systems and features like Community Notes on X as positive models for addressing harmful content online, and others recommended child-specific versions of platforms, similar to YouTube Kids, to create safer spaces for youth. Several participants also described solutions such as parents monitoring children’s online activity, keeping kids off platforms entirely, or encouraging people to use reporting and flagging tools. 

Still, there was a strong current of skepticism running through the responses. Some felt that “regular people can’t do anything” or that the internet is simply too vast to control. Others doubted platforms would ever solve the problem, since their business model depends on engagement and profit. 

Ana E., Spanish-dominant, 57, Democrat, Colorado: “Pues no lo sé… a lo mejor es un sueño, ¿verdad? [solucionar este problema] Pero yo pienso que sí, que las plataformas tendrían que poner más restricciones o más requisitos para que una persona pueda compartir cierto tipo de contenido en redes sociales (...) Creo que debería haber más restricciones, reglas de privacidad o seguridad para que las personas creen contenido más verídico. Aunque digo que es como un sueño, porque probablemente no sea así; es casi imposible que una plataforma esté al pendiente de todo lo que se genera, porque es muchísimo contenido.”

PJ T., English-dominant, 27, Independent, Florida: “It really comes down to the parents. When I was a kid, it didn’t matter what was on the Internet (...) Free will is a big thing: people should be able to access different content as long as it’s not inciting violence. Even if someone has a controversial opinion, it should be out there so others can see it and challenge it. But when it comes to kids, it falls on the parents. They have to regulate social media, regulate phones, and be involved in their kids’ lives.”

John P., Spanish-dominant, 34, Democrat, North Carolina: “Como está el factor dinero, siento que es muy difícil [solucionar]. ¿Por qué? Porque dejarían de generar dinero. Si hay dinero, se hace lo que sea con tal de ganarlo. Y como las redes sociales son una herramienta para generar dinero, al final siempre van a querer hacer dinero y permitir lo que convenga a través de eso. Es algo inherente.”

When asked whether harmful content online should be regulated more strictly, Latinos we spoke with suggested social media platforms can do more to regulate videos, ban pages, or add warning labels to harmful content or AI-generated content. Some went as far as to suggest that companies should censor videos telling people to take action that is harmful against others. As is common, some thought the companies should do nothing, and that people should be free to post what they like even when others might believe it.

Answers to the question of whether social media content should be “more controlled” often emphasized the difficult dichotomy between protecting users, especially children, online and infringing on freedom of expression. Some people noted that social media content should be limited but not 100% controlled, that freedom of expression should be guaranteed but that it should require that users be truthful, and that the use of social media by children under 18 should be controlled but not freedom of speech. 

The answers to who should be in charge for greater regulation or control of harmful included: 

• “Online mediators”

• “Platform creators”

• “The social media companies”

• “The person putting out the content”

• “Ourselves”

• “The individuals”

• “The companies themselves should be hiring people to review the content”

• “Someone who does not have bias or judgment”

In citing what more social media companies can be doing to protect users, many people did not seem deeply aware of existing measures to moderate harmful content on the platforms. 

Tech Companies’ Perceived Role in Democracy 

Main Point: Participants see tech companies as tied to democracy in three main ways: influencing voters, pushing specific narratives or agendas through algorithms, and promoting, or at times limiting, freedom of speech. 

Some mentioned partisan support, even naming figures like Zuckerberg and Trump, while others pointed to platforms’ power to amplify or suppress content. A few viewed them positively for enabling open expression, but overall tech companies are seen less as neutral platforms and more as political actors shaping democratic narratives.

Oscar G., Spanish-dominant, 21, Democrat, Texas: “Siento que las empresas tecnológicas juegan un papel muy importante porque pueden, por ejemplo, si quieren que gane un partido o una propuesta, ponen las mejores cosas, lo lanzan, y la gente lo ve. De tanto verlo, psicológicamente terminas convenciéndote de algo, especialmente cuando no conoces bien un tema.”

PJ T., English-dominant, 27, Independent, Florida: “Well, a hundred percent they do [about big tech companies playing a role in democracy]. Meta even got sued for it. They were basically pushing one side’s content more than the other, and promoting negative content about the other side on purpose. I don’t remember the exact reason, but they went to court and got in trouble for it.”

Yanina A., Spanish-dominant, 26, Republican, Texas: “[Affirming tech companies’ role in democracy] Porque ellos son quienes dan la información y hacen que las personas se informen bien o mal. Hay mucha información que es real y otra que es falsa, y la gente se asusta. Incluso, cuando pasaron muchas cosas, algunas personas cerraron negocios o se fueron a sus países.”

A few participants see tech companies’ power as appropriate or beneficial (e.g., enabling free speech), and yet others feel platforms promote democracy by enabling open expression, allowing anyone to share opinions freely.

METHODOLOGY 

“Conversations with Latinos” comprised a qualitative board of 28 participants over the course of four (4) days, followed by interviews with 14 of the participants.

The sample had the following characteristics:

Age

Total Age Range: 21–65 years old

Qual Board Participants (28):

• 20 - 30 years old: 9 participants 

• 30 - 40 years old: 8 participants 

• 40 - 50 years old: 6 participants 

• 50+ years old: 5 participants 

Interview Participants (14): 

• 20 - 30 years old: 6 participants 

• 30 - 40 years old: 6 participants 

• 40 - 50 years old: 1 participant

• 50+ years old: 1 participant

Gender

Qual BoardParticipants (28):

• Female: 14 participants

• Male: 14 participants

Interview Participants (14):

• Female: 6 participants

• Male: 8 participants 

Geographic Location

Representation across multiple U.S. states, with participation as follows:

Qual Board Participants (28):

• Texas: 4 participants 

• California: 4 participants 

• Florida: 4 participants 

• New York: 3 participants 

• Illinois: 3 participants 

• Arizona, Colorado, Georgia, North Carolina & Nevada: 2 participants each

Interview Participants (14):

• Texas: 2 participants 

• New York: 2 participants 

• Florida: 2 participants 

• Illinois: 2 participants 

• North Carolina: 2 participants 

• California, Nevada, Colorado, and Arizona: 1 participant each 

Country of Origin/Heritage 

Qual Board Participants (28):

• Mexico: 10 participants

• Puerto Rico: 5 participants

• Venezuela: 3 participants

• El Salvador, Guatemala and Cuba: 2 participants each

• Panama, Colombia, Peru and Ecuador: 1 participant each

Interview Participants (14):

• Mexico: 4 participants

• Puerto Rico: 2 participants

• Guatemala: 2 participants

• El Salvador, Cuba, Colombia, Peru, Ecuador and Venezuela: 1 participant each

Language Preference - Spanish/English

Qual Board Participants (28):

• English-dominant: 11 participants

• Spanish-dominant: 17 participants

Interview Participants (14): 

• English-dominant: 7 participants

• Spanish-dominant: 7 participants

Political Affiliation

Qual Board Participants (28):

• Democrat: 11 participants

• Republican: 9 participants 

• Independent: 8 participants

Interview Participants (14):

• Democrat: 6 participants

• Republican: 4 participants 

• Independent: 4 participants

ACKNOWLEDGMENTS

The report authors would like to thank the team at Castillo & Associates for their partnership and support on this research project. We would also like to thank Laura Zommer, Ana María Carrano, and Olivia Rivarola at Factchequeado for joining us in making this study a reality, and for their commitment to a shared mission of creating a healthier online world for Latino communities in the United States. 

A big thank you to Roshni Nedungadi of HIT Strategies for her expert insights and contributions as an advisor on this work, and to Caitlin Opperman of Equis Research for offering us feedback throughout this process. 

As always, DDIA thanks Donald Partyka for his creativity and expertise in making the design of this report. 

APPENDIX

As part of this study, one of the four days of Qual Boards were dedicated to engagement and reactions to three videos produced by Factchequeado. The three videos address the same topic — a scam involving a message that claims to be from USPS — but each uses a different format and storytelling approach: 

1. Video 1: Stream Format (This video uses a stream-style format, with a person speaking directly to the camera and a visible microphone, as if recording a podcast, in a friendly, conversational tone.)
   

2. Video 2: Host on camera with dynamic video background (A young female host speaks on camera in front of a dynamic background with fast-changing visuals related to the topic. The video is agile and set to background music.)
   

3. Video 3: Video made with AI-generated visuals and voiceover (The video features an AI-generated male voiceover, with background visuals made up mostly of stock images, illustrations, and some videos related to the topic, accompanied by background music.)

Download the Report:

SPANISH TAKEAWAYS

CONVERSACIONES CON LATINOS:

¿Qué hace que un contenido te atrape? Cómo los latinos en Estados Unidos navegan por el ecosistema de información digital actual

Introducción y contexto

Las redes sociales y las aplicaciones de mensajería se han arraigado profundamente en nuestra vida cotidiana. Desde el momento en que abrimos los ojos por la mañana, estamos constantemente con nuestros teléfonos, tabletas o computadoras. Sin embargo, aunque el estadounidense promedio consulta sus dispositivos móviles más de 160 veces al día, se sabe relativamente poco sobre qué hace que el contenido llame la atención mientras estamos conectados, qué resuena en qué audiencias y qué es lo que mucha gente realmente desea en Internet, tanto consciente como inconscientemente.

Comprender los factores psicológicos, sociales y de consumo de medios que impulsan la interacción con la información en línea es el objetivo principal de la misión del Digital Democracy Institute of the Americas (DDIA) para fortalecer una Internet más saludable para la democracia. A partir de tres encuestas fundamentales y una guía de mejores prácticas para crear contenido de prebunking (desmentido preventivo) para los latinos, en 2025 el DDIA se propuso escuchar más profundamente a los latinos sobre sus experiencias en línea, en las redes sociales y en las aplicaciones de mensajería.

En colaboración con la empresa de investigación Castillo & Associates y la organización latina de verificación de datos Factchequeado, con el apoyo de Roshni Nedungadi, miembro del Consejo Asesor del DDIA y cofundadora de HIT Strategies, durante los meses de julio y agosto de 2025, el DDIA llevó a cabo «Conversaciones con latinos», una serie de estudios a partir de paneles cualitativos (qual boards) y entrevistas destinados a comprender mejor lo siguiente:

1. Cómo consumen los latinos estadounidenses los contenidos en las redes sociales y qué es lo que «tiene» ese contenido que capta su atención. 

2. En quién confían más los latinos estadounidenses en las redes sociales y cómo eso se relaciona o no con lo que eligen para entretenerse y mantenerse informados. Para evaluar esto, se analizaron las características de las cuentas fiables (señales positivas) y lo que hace que la gente se muestre escéptica ante determinadas cuentas (señales negativas). 

3. Si los participantes reconocen, interpretan y verifican la información engañosa o falsa cuando se exponen a ella, y de qué manera lo hacen.

4. Qué perciben los latinos como los mayores retos para la democracia y la sociedad.

5. Cómo perciben los latinos el uso de la IA en la creación o influencia de la información, sus riesgos y beneficios, y cuáles son sus opiniones sobre el contenido generado por la IA. 

6. Las percepciones de los latinos sobre el papel de las grandes tecnológicas en la democracia, la regulación de las plataformas, la censura, los algoritmos y la seguridad en línea, con especial atención a cómo las personas experimentan los entornos digitales.

Las conclusiones descritas en las cinco secciones de este informe pueden utilizarse para dar forma a lo siguiente:

• Información relevante para los influencers y creadores latinos comprometidos con la creación de contenidos saludables y basados en hechos en Internet. 

• Intervenciones y estrategias de desarrollo de capacidades para mejorar la producción de contenidos, la credibilidad y la competencia cultural.

• Programas de formación para periodistas, creadores de contenidos, científicos, responsables políticos y otros comunicadores de primera línea que trabajan para llegar a los latinos.

• Políticas relacionadas con las redes sociales, la confianza y la seguridad, y la gobernanza de Internet. 

«Conversaciones con latinos» consistió en un panel cualitativo (una herramienta de investigación cualitativa asincrónica en línea similar a un grupo focal en línea) de 28 participantes, seguido de entrevistas en profundidad con 14 de ellos. 

La muestra incluía una mezcla casi equitativa de latinos con dominio del inglés y del español, mujeres y hombres, de diversos grupos de edad y afiliaciones políticas. El grupo estaba compuesto por latinos que vivían en los estados de Arizona, California, Colorado, Florida, Illinois, Nueva York, Carolina del Norte, Nevada y Texas. (Al final del documento se incluye un desglose completo de la metodología y los datos demográficos de los participantes). 

Este estudio se diseñó para complementar los datos de la encuesta DDIA existente, así como otros estudios en este campo, añadiendo matices y revelando momentos clave detrás de los comportamientos digitales de los latinos en Estados Unidos. 

Al tratarse de un estudio cualitativo, los resultados descritos no pretenden ser representativos de todos los latinos en Estados Unidos, sino ilustrar ideas y motivaciones que pueden ser difíciles de captar a través de la investigación cuantitativa habitual. 

Puntos clave por sección 

Puntos clave - SECCIÓN 1: Hábitos digitales, consumo de contenidos y fuentes de referencia 

CONTENIDO QUE LLAMA LA ATENCIÓN 

1. DESPLAZAMIENTO PASIVO VS BÚSQUEDA ACTIVA. Cuando se les preguntó qué contenido les llamaba la atención, la mayoría de los participantes mencionaron primero el «tema» de las publicaciones. Al insistirles, muchos mencionaron el encuadre del contenido y el comportamiento de los emisores. Cabe destacar que lo que llama la atención y se valora cuando las personas se desplazan pasivamente por sus feeds en redes parece diferente de lo que llama la atención y se valora cuando las personas buscan activamente (véanse los detalles a continuación). 

DESPLAZAMIENTO PASIVO

2. TEMAS DE INTERÉS. Los latinos con los que hablamos valoraban el contenido que trataba temas que les interesaban personalmente, lo que comprendía principalmente una mezcla de contenido divertido y ligero y contenido informativo (definido aquí como contenido que enseñaba a las personas algo nuevo). 

3. CONTENIDO DESEADO. Las personas anhelan positividad y autenticidad en las redes sociales. El contenido alegre y/o informativo (crecimiento personal) y/o centrado en la comunidad destacó entre los participantes como atractivo y deseable. 

4. LAS PERSONAS COMO SUJETOS. Los amigos, famosos o influencers como sujetos del contenido suelen llamar la atención. Los participantes con los que hablamos indicaron que dejan de desplazarse por el contenido de sus feeds cuando aparece alguien que reconocen. 

5. EMISORES. Muchos latinos expresan que valoran más la interacción con cuentas individuales que con cuentas de organizaciones cuando se desplazan pasivamente por la pantalla. Entre los comportamientos y características de las cuentas individuales que se valoraron se incluyen el compartir evidencias, la autenticidad, la proximidad y la accesibilidad (más información sobre lo que indica esto en el texto principal).

6. FORMATO. Muchas personas valoran las cuentas individuales que presentan la información de forma «limpia», directa y bien producida. Muchos se inclinan por videos o imágenes centrados en las personas. 

BÚSQUEDA ACTIVA 

7. COMPORTAMIENTOS DE BÚSQUEDA DE INFORMACIÓN. En la búsqueda activa, los participantes describieron comportamientos de búsqueda activos o híbridos cuando desean informarse: la mayoría combina el desplazamiento en el feed de contenido con búsquedas de seguimiento o busca información directamente. Las redes sociales, especialmente Instagram y TikTok, son puntos de entrada comunes para contenido informativo, y cuando la información específica despierta el interés de las personas, estas indican que acuden a Google, YouTube o ChatGPT para verificar la información u obtener más contexto.

8. CONFIANZA. Aunque la mayoría de los latinos con los que hablamos señalaron que no confían al 100% en ninguna cuenta de redes sociales, ya sea de personas u organizaciones –mencionando elementos como el sesgo, la corrupción y el interés propio–, el elemento de la confianza sigue siendo importante para las personas que buscan información de forma activa. La confianza de los participantes en las fuentes durante la búsqueda activa parece depender del contexto, basándose en el tema, la fuente, el contenido y el comportamiento de quienes transmiten el mensaje (los emisores). YouTube se perfiló como la plataforma de redes sociales más confiable. 

9. SEÑALES DE CONFIANZA. Muchos de los latinos con los que hablamos indicaron que se apoyaban en ciertas señales para determinar la confiabilidad y credibilidad de la información y de las cuentas. Estas incluían la presentación de pruebas para respaldar opiniones, las insignias de verificación, la popularidad de las cuentas, la redacción de los comentarios, las funciones de suscripción voluntaria y un historial de precisión en el pasado. 

10. ELEMENTOS PERSUASIVOS DEL CONTENIDO. Las perspectivas de primera mano, como las entrevistas directas y las grabaciones en el terreno, aumentaban el carácter persuasivo del contenido. Algunas personas también señalaron que ver puntos de vista contrastantes, incluyendo diferentes lados de una historia, les llevaba a replantearse sus perspectivas sobre ciertos temas. 

11. EL PAPEL DE LA REPRESENTACIÓN. Los latinos con los que hablamos notaron que ciertos rasgos influían en cómo percibían si un emisor los representaba de manera personal. Estos rasgos iban desde las conexiones culturales o comunitarias, hasta la defensa de la inmigración u otros temas importantes para ellos, pasando por el estilo de comunicación del emisor.

Puntos clave - SECCIÓN 2: Daños en línea

12. AMENAZAS A LA DEMOCRACIA. Cuando se les pidió que nombraran las mayores amenazas a la democracia en la actualidad, casi todos los 28 participantes del panel cualitativo citaron alguna forma de  daños en línea en el entorno de la información digital en la difusión de contenidos falsos o engañosos que socavan la confianza. Algunos también mencionaron el uso político del poder, la percepción de que los líderes actúan en su propio interés, se extralimitan o fomentan la polarización en lugar de representar al pueblo. También se mencionaron otros temas como la injusticia racial, la libertad de expresión y la política exterior. 

13. MIEDO O INCOMODIDAD A LA HORA DE EXPRESARSE. Los participantes compartieron con frecuencia que evitan expresar sus opiniones en línea por miedo al odio, a las reacciones negativas o a la confrontación. 

14. POLARIZACIÓN. La política, las elecciones y la inmigración surgieron como los temas polarizantes más citados. Los participantes describieron la sensación de que, en lo que respecta a estos temas, los espacios en línea dejaban poco margen para los matices o las posiciones intermedias. 

15. CONTENIDO DAÑINO. Cuando se les pidió que dieran su opinión sobre qué contenido consideraban más dañino en internet, los encuestados mencionaron con frecuencia el contenido falso o engañoso, el discurso de odio, el acoso en línea y el contenido que fomenta el miedo. Las narrativas sobre inmigración y la retórica racista se consideraban contenidos dañinos frecuentes dirigidos específicamente a los latinos. 

16. MOTORES DE DESINFORMACIÓN. Las motivaciones más mencionadas en la difusión de desinformación fueron: creadores de contenido que buscan «me gusta», dinero o viralidad; actores políticos que difunden el miedo o moldean narrativas para obtener beneficios electorales; y motivaciones relacionadas con el racismo o a la dominación sobre los latinos. 

17. TEMA DESINFORMANTE MÁS FRECUENTE: LA INMIGRACIÓN. Todos los participantes afirmaron que ellos mismos, o alguien de su familia o círculo cercano de amigos, habían estado expuestos a la desinformación en algún momento. La desinformación relacionada con la inmigración, incluida la información falsa sobre redadas, restricciones y medidas coercitivas, se reveló como la más común en la actualidad. También se mencionaron con frecuencia las estafas en las aplicaciones de mensajería. 

18. VERIFICACIÓN DE HECHOS. Las personas indicaron que compartirían verificaciones de hechos principalmente cuando la información les afecte personalmente a ellas o a las personas que las rodean. La mayoría de las personas mencionaron que compartirían las verificaciones directamente con amigos y familiares, en lugar de publicarlas en las redes sociales. Algunas personas, cuando sospechan que se trata de «noticias falsas», lo denuncian directamente a las plataformas, pero muchas no entienden que esa es una opción.

Puntos clave - SECCIÓN 3: Inteligencia Artificial (IA)

1. PERCEPCIONES GENERALES SOBRE LA IA. Muchos encuestados tienen sentimientos encontrados sobre la IA, aunque bastantes señalaron que la IA les ha facilitado el día a día. Si bien muchos consideran que las herramientas de IA son neutras, señalan que la IA puede ser positiva o negativa según quién la utilice y con qué propósito. Algunos participantes también expresaron incertidumbre sobre la presencia de la IA en sus vidas, reconociendo que «probablemente está justo delante de sus narices» de formas que quizá no perciben o comprenden del todo. 

2. PERCEPCIONES DE LA IA BASADAS EN LAS EXPECTATIVAS. La forma en que las personas perciben el contenido generado por IA cuando lo encuentran en internet o en las redes sociales se ajusta a sus expectativas sobre cómo interactúan con la información y en qué espacios. Al desplazarse pasivamente por la pantalla, los encuestados indicaron que se sentían «molestos» al encontrar contenido de baja calidad generado por IA, mientras que, al buscar activamente, muchos confían en que la IA les proporcione contenido fiable e imparcial. 

3. GRADO DE EXPOSICIÓN AL CONTENIDO GENERADO POR IA EN INTERNET. Las conversaciones con los participantes revelaron que las personas piensan que se encuentran con contenido generado por IA con mucha más frecuencia que en los últimos dos años. 

4. CAPACIDAD PARA DETECTAR LA IA. En general, los participantes con los que hablamos creen saber cuándo el contenido que ven ha sido generado por IA. Utilizan principalmente señales visuales y de voz para evaluar si el contenido ha sido generado por IA.

5. PERCEPCIÓN SOBRE LA VULNERABILIDAD DE LOS DEMÁS. Aunque muchas de las personas con las que hablamos pensaban que podían detectar la IA cuando la veían, muchos participantes no creían que otras personas, especialmente los adultos mayores o los niños, pudieran distinguir el contenido generado por IA del contenido «real». 

6. ETIQUETADO DEL CONTENIDO GENERADO POR IA. En casi todas las 14 entrevistas, cuando se les preguntó si el contenido generado por IA debería etiquetarse, la mayoría de los encuestados indicaron que creían que sí. Los participantes señalaron que las plataformas deberían marcar el contenido generado por IA para «evitar confusiones».

Conclusiones principales - SECCIÓN 4: Las grandes tecnológicas y la normativa

1. QUIÉN DECIDE LO QUE VEMOS. La mayoría de los participantes reconocen que los algoritmos de las plataformas son un factor decisivo para lo que vemos en nuestros feeds, y varios señalan que las plataformas promocionan intencionadamente contenidos que generan un alto nivel de interacción, lo que muchos reconocen que puede dar lugar a la difusión de contenidos nocivos. Se repite la creencia de que «los teléfonos nos escuchan» y se reconoce que esto alimenta los anuncios o contenidos relacionados. 

2. RIESGOS GENERALES EN INTERNET. Cuando se les preguntó cuáles eran los mayores riesgos percibidos en internet, los participantes del estudio señalaron desde la desinformación y la promoción de la violencia en línea, hasta la privacidad y la vigilancia, pasando por los hackeos y las estafas, y el acoso selectivo. Muchos participantes mencionaron el temor a ser rastreados en internet o a que les robaran su información o datos personales. 

3. VULNERABILIDAD A LOS RIESGOS EN LÍNEA. Al igual que en el caso del contenido generado por IA en la sección anterior, los encuestados en general se consideraban capaces de detectar contenido dañino en línea. Los grupos que muchos participantes consideraban más «vulnerables» a los daños en línea incluían a los niños y los adultos mayores, y algunas personas compartieron que creen que todas las personas en línea son vulnerables a los riesgos en algún momento.

4. ¿QUIÉN ES RESPONSABLE DE GARANTIZAR LA SEGURIDAD DE NUESTRO MUNDO ONLINE? La mayoría de los latinos con los que hablamos atribuyeron la responsabilidad de garantizar un mundo online saludable a las plataformas de redes sociales y a las empresas tecnológicas, sugiriendo que podrían crear filtros más potentes, prohibir cuentas perjudiciales, hacer cumplir los derechos de autor e introducir procesos de revisión más estrictos antes de mostrar el contenido. Muchos parecían desconocer hasta qué punto las empresas ya hacen algunas de estas cosas. Varios participantes hicieron hincapié en que las personas y los padres también desempeñan un papel crucial en la seguridad de Internet. Algunos señalaron la importancia de que los usuarios sean «auténticos» en línea y no se limiten a buscar «me gusta», planteando parte de la solución como una cuestión de responsabilidad personal. 

5. EL PAPEL DE LAS EMPRESAS TECNOLÓGICAS EN LA DEMOCRACIA. Los participantes consideran que las empresas tecnológicas están vinculadas a la democracia de tres maneras principales: influyendo en los votantes, impulsando narrativas o agendas específicas a través de algoritmos y promoviendo (o, en ocasiones, limitando) la libertad de expresión.