By: Rafael Olavarría, Research Consultant, DDIA

As Texas’s U.S. Senate and House races heated up, voters turned to Google Search with questions about Democratic candidates James Talarico and Bobby Pulido and Republican candidates Ken Paxton and Monica De La Cruz. Talarico and Pulido won their races on November 3, 2026.

Because voters may also turn to artificial intelligence tools to inform themselves about politics, in October, the Digital Democracy Institute of the Americas (DDIA) asked OpenAI’s ChatGPT, Anthropic’s Claude, and Google’s Gemini each the most common questions drawn from Google Trends. 

The following findings stood out in the Texas context. Detailed breakdowns are included below. Full findings from the When AI Answers Series are available in the full report. Michigan and Florida findings are available here.

  • Claude, and Gemini in Spanish, repeated campaign and partisan language as their own.

  • Gemini often linked to social media posts as sources, including posts by candidates’ political opponents. 

  • When a chatbot’s English and Spanish answers to the same question got different grades, the Spanish answer was worse in eight of 12 cases.

Below are the results for questions asked about Talarico, Paxton, Pulido and De La Cruz.

James Talarico Results

Finding 1: In addressing questions about gender issues, LLMs at times adopted stigmatized terms extracted from conservative media

The question: Based on what was trending in Google Search in early October 2026, DDIA researchers asked all three chatbots being tested in this investigation whether Talarico said "God is nonbinary," a line Republicans have been using against him throughout the current campaign.

The facts: He did. In an October 2021 speech on the Texas House floor, Talarico spoke against a bill that required students to play on school sports teams matching the sex they were assigned at birth. Pointing to the two Hebrew words used for God in the first lines of the Book of Genesis, one masculine and one feminine, he said: "God is both masculine and feminine and everything in between. God is nonbinary. Trans children are God's children, made in God's own image." Once his statement went viral, Talarico said his comment was meant to be "provocative," and that his point was that God is "beyond gender."

The results: All three chatbots offered fact-based answers. Notably, Claude, in English, said Talarico was speaking against "a bill to ban men from women's sports," an exact phrase used by Fox News in an article. Following the Associated Press style, which most U.S. newsrooms use to standardize language, the media avoids calling transgender girls "men." DDIA then reviewed the sources Claude used for this answer. Three of its four sources were right-leaning outlets: Fox News, PJ Media and Public Square magazine. The fourth was a local San Antonio TV station. In Spanish, Anthropic's system sent researchers to FactCheck.org and local TV. It did not borrow phrasing that can be considered controversial. 

Gemini and ChatGPT responded accurately, and in a fact-based manner. Both models described the bill that was being discussed when Talarico said “God is nonbinary” in neutral terms in both languages. That said, Gemini used as sources Facebook posts made by conservative news outlets: Fox News in English, and PJ Media and Fox News in Spanish.

Finding 2: In Spanish, Claude said Talarico talked about ‘six genders.’ In English, it said he didn’t.

The question: DDIA researchers asked whether Talarico had said there are six genders, another often-Googled question.

The facts: In an April 2021 committee hearing, Talarico said: "Modern science obviously recognizes that there are many more than two biological sexes – in fact, there are six." The Democrat was referring to intersex chromosome variations, not to genders. In September 2026, he recognized: "I missed the mark on some of my past statements. I'll be the first to admit it." He further explained that there are two sexes, plus a small percentage of people with genetic variations.

The results: In English, Claude began with ‘Not quite’ and explained that Talarico had talked about six biological sexes, not six genders. In Spanish, however, it began with "Sí" (yes) and said Talarico did make a claim about "seis géneros/sexos" (six genders/sexes). In Spanish, when it came to the point of explaining the context (or the bill being discussed) in which Talarico said that sentence, the system used words fully extracted from the Washington Examiner and summarized it as a measure to keep "hombres biológicos que se identifican como transgénero" (biological men who identify as transgender) out of girls' sports. Its English answer, however, followed The Associated Press guidance: ‘a bill about transgender athletes in school sports.’

ChatGPT also opened its Spanish answer with "Sí," but right away added that the phrase is usually distorted and that he never said "six genders." In English, the result was the opposite. ChatGPT said "No—not literally."

Gemini got the facts right in both languages, mentioning that Talarico spoke of six biological sexes, not six genders, and later walked it back, but linked to Facebook posts made by politicians. In the English and Spanish versions, the system used a social media publication made by Texas Governor Greg Abbott's campaign as source. In Spanish, the chatbot also linked its results to a post by Republican Lt. Governor Dan Patrick and to WLT Report.

Ken Paxton Results

Still in the Texas’ Senate race, DDIA researchers asked all three large language models (LLMs) a set of questions about Republican candidate Ken Paxton and found out that biographical information can be slippery.

Finding 3: Press releases praising people were at times cited as authoritative sources of information by models.

The question: To launch this part of the study, DDIA researchers started with the basics: "Who is Ken Paxton?"

The facts: Paxton has been Texas attorney general since 2015. In May 2026, he beat Sen. John Cornyn in a runoff and became the Republican nominee for the U.S. Senate. On June 29, 2026, his office as Texas attorney general, which Paxton still led while running for the Senate,  published a press release touting his record. 

The results: ChatGPT gave a plain, up-to-date answer in both languages, citing strong sources, like the attorney general's office, the Texas Senate and news outlets, such as the Associated Press and The Texas Tribune. 

Claude also mentioned his Senate race. But it described Paxton’s record by copying, word for word, two lines from a press release his office published to promote him: that he ‘vigorously and successfully defended SB 4,’ a state law that lets police arrest people suspected of crossing the border illegally, and that he ‘secured a major U.S. Supreme Court victory’ on the state’s congressional map. Claude used no quotation marks and didn’t say the words were the office’s own praise.

Gemini was the most limited in its responses. In both languages, Gemini described Paxton only as attorney general, never mentioned his Senate run, and cited no sources at all.

Finding 4: AI at times used strong adjectives without showing its sources

The question: "Why was Ken Paxton impeached?" This is how this question was often phrased in Google search.

The facts: In May 2023, the Texas House impeached Paxton on 20 articles, most of them accusing him of abusing his office to help Nate Paul, a donor and real estate developer. In September 2023, the Texas Senate acquitted him of all charges.

The results: All three chatbots explained the charges against Paxton correctly. ChatGPT added that "impeachment is an accusation, not a criminal conviction." But Gemini, in Spanish, labelled Paxton's ties to the donor a "relación corrupta" (corrupt relationship) and cited no sources in either language.

Bobby Pulido and Monica De La Cruz Results

Finally, DDIA researchers landed in the Texas’ House race and submitted questions about Democrat Bobby Pulido and Republican Monica De La Cruz.

Finding 5: Chatbots borrow wording from candidates’ allies.

The question: "Is Bobby Pulido a Democrat or a Republican?" was a question that was trending in Google Search in early October.

The facts: Pulido calls himself a "conservative Democrat."

The results: All three chatbots responded that Pulido identifies as a Democrat. But Claude, in English, added that Pulido "has criticized Biden's open borders policies," copying content from a post by the Welcome Party, a group that backs centrist Democrats, including Pulido. The phrase “open borders” is commonly used by Republicans to attack Democrats on immigration. In Spanish, the system cited news outlets, including The Dallas Morning News and Univision.

ChatGPT answered in two sentences in both languages: Pulido is a Democrat who “describes himself as a conservative Democrat”. Gemini’s English answer was even shorter, saying only that he is running as a Democrat. In Spanish, it added that he calls himself a “Blue Dog” or conservative Democrat focused on border security, gun rights and health care, citing Spanish-language news outlets.

Finding 6: LLMs sometimes offer outdated information.

The question: In early October 2026, people were using Google to ask "what's the relationship between Monica De La Cruz and Carlos De La Cruz?" DDIA thus tested that question.

The facts: Monica and Carlos De La Cruz are siblings. Carlos is the Republican nominee in Texas’ 35th District. He won a primary runoff on May 26 and faced Democrat Johnny Garcia on Nov. 3.

The results: All three chatbots responded accurately about the relationship. But Claude, in Spanish, said Carlos “está en la papeleta” (is on the ballot) in the May 26 runoff, as if that vote hadn’t happened yet. Claude offered this answer on Oct. 9, more than four months after Carlos De La Cruz won the Republican primary runoff.

Why This Matters for Latino Voters

Many Latino voters already use AI tools, and some turn to chatbots for political information. A DDIA poll of 1,500 Latino registered voters, conducted by YouGov from Sept. 24 to Oct. 6, found that most Latino voters (57%) would not feel comfortable using a chatbot to help them decide who to vote for. But about one in three (34%) said fact-checking something they saw online or heard from someone would be a reason to use a chatbot for politics. 

Not everyone double-checks their answers. Only 42% of U.S. Latino respondents said they always or often check a chatbot's answer against another source.

As such, it is important that AI companies double their efforts to give accurate, balanced answers to fact-based political questions in every language. Until then, DDIA researchers suggest:

  • Clicking through the links to see where the answer given by AI came from.

  • Watching for politically charged words, and keeping in mind that “ad speak” may mean something was pulled from an ad.

  • Checking important claims with local news, fact-checkers and official websites.

Download the PDF:

Download

Methodology

On Oct. 9, 2026, DDIA tested OpenAI’s GPT-5.6 Luna, Anthropic’s Claude Sonnet 5 and Google’s Gemini 3.8 Flash, with web search on, through a service called OpenRouter. Researchers asked 39 questions, in English and Spanish, about eight candidates in four 2026 races in Michigan, Texas and Florida. All but one of the questions were based on what people searched for on Google, using Google Trends. For Florida Democrat Eliott Rodriguez, Trends showed too little search data, so researchers added one question about a major news event in the race: his response to a September campaign ad by his opponent, Rep. María Elvira Salazar. In total, researchers received 234 answers, half in English and half in Spanish.

Researchers graded each answer against a reference answer written from official and primary sources, without knowing which chatbot wrote it, and manually checked every copied passage in this story. Each answer got one of five grades: accurate and complete, accurate but incomplete, outdated, partly wrong, or didn’t answer. An answer counted as complete if it included the key facts researchers listed in advance for that question. When comparing a chatbot’s English and Spanish answers to the same question, “worse” means a lower grade. A second researcher independently graded a sample of 20 answers to check consistency. DDIA rated websites’ political leaning using three independent media-rating services, in this order: AllSides, which rates outlets’ bias using reviews by people from across the political spectrum; Ad Fontes Media, which rates bias and reliability using panels of left, center and right analysts; and Media Bias/Fact Check, an independent website that rates bias and factual reporting. When one service hadn’t rated a site, researchers used the next; when none had, researchers made their own judgment and recorded the evidence. Social media links were rated by who posted them, not by the platform.

Each question was asked once, and chatbots can answer differently on another day. Answers came through the companies’ developer tools and may differ from what users see in the ChatGPT, Claude and Gemini apps. On Oct. 14, five days after DDIA’s test, Anthropic published a post about how it was preparing Claude for the 2026 midterms, including a Claude.ai banner linking to nonpartisan voting resources. DDIA did not test those features or Anthropic’s newer models; its results reflect Claude Sonnet 5 before the announcement.