As candidates utilize AI tools and deepfake versions of political rivals, the line between satire and disinformation blurs, leaving voters concerned.
A study conducted by the London School of Economics found that nearly 40% of UK respondents found the use of AI-generated materials for political campaigns to be unacceptable.
Currently, Texas allows deepfakes in political campaigns unless they are created and distributed within 30 days of an election, according to the National Conference of State Legislatures.
In Texas, the decision to require a disclosure on certain political advertising and the banning of using certain artificially generated media is pending, as outlined by the NCSL.
Anthony Yañez, a Texas Tech law student and president of OUTlaw, a LGBTQIA+ organization for Texas Tech law students, said he doesn’t like generative AI whether it's in political campaigns or otherwise.
“It just feeds into this idea of laziness. If you’re going to be my representation, my person in the office, why are you putting out an AI ad that shows your laziness? It tells me they’re going to be lazy on the job too,” said Yañez.
Recently, both Lieutenant Governor Dan Patrick and Attorney General Ken Paxton have received criticism across social media platforms, like Facebook and Instagram, for their use of deepfakes in political advertisements.
Patrick's campaign released ads featuring Representative Vikki Goodwin that she alleged were deepfakes that misrepresented her views to sway the election, according to Houston Public Media.
Patrick's ads resulted in a police report, filed by Goodwin, and have since been taken down.
Paxton's ad featured an AI-generated version of Representative James Talarico writing "let there be six genders' in the bible, calling God nonbinary, and calling himself a 'Christian who hates Christianity,' among other quotes that have been taken out of their original context.
The comments on Paxton's ad on YouTube are largely negative, with one commenter comparing it to the show Black Mirror.
Another commenter expressed that they believed only the elderly would believe these deepfake advertisements.
However, a study conducted by the University College London found that humans cannot reliably detect speech deepfakes, regardless of the language spoken.
Additionally, a joint study between Neuro-Com Research Group, RTVE Institute and Universitat Autònoma de Barcelona found that participants wrongly identified nearly 39% of AI-generated material as being human-made.
"If the public realizes that an ad is fake, they will take it out on the originator. Americans do not like to be duped," said Anthony Lastrape, a professor who taught political campaign communication at Texas Tech University, in an email.