Ain’t gonna happen this year, though.
Artificial intelligence and prediction markets have opened new frontiers in Texas politics, but one expert said the growth of platforms like ChatGPT and Polymarket is far outpacing legislative efforts to limit their impact on elections.
It’s a trend that’s emerging in Texas’ Republican primary for U.S. Senate, as Sen. John Cornyn and Attorney General Ken Paxton trade AI-generated attack ads, including a recent video published by Paxton that featured a smiling Cornyn enjoying “spring break” while President Donald Trump’s agenda stalled in the Senate.
The race has also seen significant interest from traders on prediction markets, which allow users to bet on the outcome of everything from elections to wars and sports. More than $15 million in trades have been placed on the outcome of the primary through Polymarket, according to the website’s market page for the Senate primary.
Texas was the first state to pass a law limiting the use of deepfakes in political advertising, but experts said the technology has evolved well beyond the scope of the 2019 bill, which applies only to state races. Further state and federal efforts to regulate the impact prediction markets and other AI tools can have on elections have largely floundered, leaving the door open for political actors to use the technologies in everything from the 2024 presidential primary to a 2025 election for Dallas City Council.
Andrew Cates, a campaign ethics attorney who’s written nine editions of a legal practice guide on the subject, said, without guardrails, both technologies can have dire consequences for elections in Texas.
“We’re cooked for sure,” Cates said. “Nobody knows anything about AI (at the Capitol). Nobody knows anything about computers. And that’s beyond the fact that the absolute truism is that government will never keep up with technology. It moves way too fast.”
The information ecosystem of elections — already burdened by disinformation on social media — could collapse as candidates adopt increasingly advanced tools that make it near-impossible for voters to distinguish fact from fiction, Cates said.
Voices can be spoofed, videos and images fabricated. Trend lines on prediction markets can be driven up or down by deep-pocketed candidates and donors, potentially impacting voters’ perceptions of candidates’ popular support, Cates said.
“You can just have somebody be a straw man and dump money into a Polymarket race and influence everybody else,” Cates said. “Anyone can drop a bunch of money and then go: ‘Oh, s—, somebody just put $1 million on Donald Trump.’ And then all of a sudden, it swings it up to a 70-30 race.”
Cates said Texas law currently does not prevent candidates from using their campaign donations to bet on themselves using prediction markets like Polymarket, which did not respond to questions about the company’s policies regarding elections.
[…]
Cates worked with Rep. Dade Phelan to introduce a bill building on the 2019 deepfake law, hoping to force candidates and PACs to disclose if AI was used in political advertising. The bill passed the House but died in the Senate.
Phelan, who previously served as speaker of the Texas House, announced in August he would not seek reelection. The Beaumont Republican’s decision came amid rising ire from GOP voters who have criticized him for his role overseeing Paxton’s 2023 impeachment proceedings and attacks on his record as a conservative lawmaker.
Phelan said political social media influencers were among the most vocal critics of the bill, saying it was an attack on political speech and their ability to “meme” freely.
But Phelan said the bill aimed only to regulate memes that were generated by people being paid to do so. Regulating political speech funded through campaign contributions is a common practice, he said, pointing to disclosure requirements for political advertisements.
“There was a segment of paid online influencers who probably would have had to disclose their AI use under this bill, and they’re the ones who try to say it had to do with memes,” Phelan said.
[…]
Phelan said he’s also concerned about prediction markets and the impact they may have on Texas elections. But without action from the federal government, he said he doesn’t expect to see legislation addressing AI or prediction markets out of Austin.
“It’s going to take, unfortunately, a very bad incident where someone loses an election over a fake ad before we see anything real happen here,” Phelan said. “It’s sad, because the voters at the end of the day are the ones that are going to lose out the most.”
Well, Grampa Greg Abbott keeps falling for AI-generated fake videos, so the fear is real. That said, I don’t think we’re cooked, not yet anyway. I do agree with Dade Phelan that it’s going to take a very bad incident for anything to happen. I think it’s also going to require a non-corrupt Supreme Court, which may have a more nuanced view of free speech – I don’t think it’s ridiculous to say that chatbots and LLMs and the like don’t have First Amendment protection, at least not as broadly as the rest of us do – and a willingness to allow normal regulation of paid advertising. Maybe some states can take the lead in the meantime, assuming Trump and the Republicans don’t manage to pass some kind of bill forbidding state action. One way or the other, something will happen. It will be a matter of cleaning it all up afterwards.
If corporations are people, then probably bots are too.