If you’ve watched television in Texas lately, you might have seen a commercial featuring Democratic U.S. Senate candidate James Talarico talking about God, gender identity, and Christianity. In one ad, Talarico appears to read from a Bible bearing his own name.
The commercial is powerful. It’s provocative. And it looks remarkably real. There’s just one problem. The person speaking in the video isn’t actually Talarico. He’s an artificial intelligence-generated version of the candidate.
The ad, released by Republican Senate candidate Ken Paxton’s campaign in September, uses AI to recreate Talarico’s likeness and voice. According to FactCheck.org, the commercial combines some statements Talarico actually made, removes important context, and includes language he never used. A disclaimer acknowledging the use of AI appears in small print near the bottom of the screen.
Talarico accused his opponent of deliberately misleading voters.
“When you look at this ad, they are using AI in order to mislead the people of this state,” Talarico told CBS News in September.
The commercial is one example of a growing problem confronting voters ahead of the Nov. 3 midterm elections. Artificial intelligence has made it easier than ever to create convincing images, videos, and recordings of politicians saying or doing things that never happened. Meanwhile, false information about voting dates, polling locations, voter registration, and election results continues to circulate on social media, via text messages, and in private group chats.
And with early voting in Texas now underway, election officials and voting rights advocates are urging voters to be especially careful about what they believe and share.
The consequences can go beyond influencing how someone votes. A misleading post about a polling location, a false warning about voter eligibility, or an inaccurate claim about identification requirements could prevent someone from voting altogether.
For Black voters, who have historically faced barriers to the ballot box, distinguishing fact from fiction is particularly important. The challenge is no longer simply recognizing a political lie. It’s recognizing one that looks and sounds like the truth.
When seeing is no longer believing
Political misinformation isn’t new. Candidates have exaggerated their opponents’ positions, circulated misleading accusations, and taken statements out of context for generations.
But artificial intelligence has changed the game. Today, someone with relatively inexpensive technology can create a video of a candidate delivering an entirely fabricated speech, generate a convincing recording of an election official announcing a false change in voting procedures, or produce a realistic photograph of an event that never happened.
And the technology keeps improving.
According to the Wesleyan Media Project, researchers identified at least 164 AI-assisted political advertisements during the 2026 election cycle, many without adequate disclosures indicating that the content was artificially generated.
Texas has already become a testing ground for the technology.
In addition to the Paxton advertisement, other political groups have circulated AI-generated videos of Talarico. PolitiFact documented advertisements that used his likeness to deliver statements he did not actually make on camera, even when portions of the underlying text came from his real social media posts.
The distinction matters. A statement can be authentic while the video supposedly documenting it is entirely fabricated.
Experts at the Brennan Center for Justice warn that voters can no longer rely on familiar signs of manipulation, such as distorted hands, unnatural facial movements, or robotic-sounding voices. AI-generated content has become sophisticated enough to avoid many of those mistakes.
Instead, voters should focus on verifying the source and context.
Who originally posted the video? Is there an authentic recording of the event? Have credible news organizations verified the statements?
A video with thousands of shares isn’t necessarily accurate. Neither is a screenshot forwarded by a trusted friend. And just because something looks real doesn’t mean it happened.
Quiz Yourself: Can You Spot Voting Misinformation?
Why misinformation spreads so quickly
One reason election misinformation is so effective is that it doesn’t always come from strangers. Sometimes it arrives in a family group chat, a church text chain, or a neighborhood Facebook group. It may come from someone voters know and trust.
That familiarity can make people less likely to question the information. A message claiming that thousands of voters have been removed from registration rolls may create panic. A video showing a politician making an outrageous statement may generate anger. A warning that polling locations have changed may prompt someone to forward the message immediately, believing they’re helping others. In each case, emotion can become more powerful than verification.
The problem is compounded when misleading content moves between platforms. A false claim may begin on an obscure website, circulate on X or TikTok, appear in a Facebook screenshot, and eventually reach voters through private text messages. By then, the original source may be impossible to identify.
The consequences also extend to election workers. A Brennan Center for Justice analysis found that 64% of election officials surveyed in 2022 said the spread of false information had made their jobs more dangerous.
Researchers and voting rights organizations have warned that misinformation can discourage participation, fuel threats against election workers, and undermine confidence in legitimate results.
The Brennan Center has also raised concerns about reduced federal resources supporting election security and efforts to counter foreign interference ahead of the 2026 midterms.
Those developments place additional responsibility on local election officials, news organizations, and voters to identify inaccurate information before it spreads.
Stop. Check. Verify.
So what should voters do when they encounter questionable election information?
Experts recommend a simple approach that doesn’t require sophisticated technology or special training.
Stop. Before forwarding a dramatic video, screenshot, or text message, consider whether the information is designed to provoke an immediate emotional reaction. Be especially cautious of messages urging people to share information before it is supposedly deleted or censored.
Check. Identify who originally posted the information. Is it from an official election office, a recognized news organization, or an anonymous account? Does the video show a complete statement, or only a short clip without context? Search for the original recording rather than relying on a repost.
Verify. Compare the claim against an authoritative source. For voting procedures, that means the Texas Secretary of State or your county election office. For political statements, look for original recordings, public records, and reporting from credible news organizations.
Voters can also consult independent fact-checking organizations such as PolitiFact, FactCheck.org, and The Associated Press.
And while AI tools can be useful for many tasks, they should not replace official election websites when determining where, when, or how to vote.
Most importantly, voters should not assume that a video is authentic simply because they cannot detect anything wrong with it.
Where Texas voters can get accurate information
The following resources can help voters independently verify election information:
| What you need | Where to verify |
| Voter registration status | Texas My Voter Portal |
| Voting dates and deadlines | VoteTexas.gov |
| Harris County polling locations | HarrisVotes.com |
| Voter ID requirements | Texas voter ID guide |
| Mail-in ballot requirements | Texas early voting FAQs |
| Harris County sample ballots | What’s on My Ballot? |
| Harris County election results | HarrisVotes.com |
| Fact-checking political claims | PolitiFact, FactCheck.org, and AP Fact Check |
Voters who encounter suspicious election information can contact the Texas Secretary of State’s Elections Division at 1-800-252-VOTE (8683).
Harris County voters can contact the county Elections Department at 713-755-6965.
Anyone who receives a suspicious message should preserve a screenshot, record the source, and report misleading information about voting procedures to the appropriate election office. Threats or intimidation at polling locations should be reported promptly to election officials or law enforcement.
