Summary
A recent political ad featuring Texas Senate candidate James Tallarico has sparked discussions about the implications of artificial intelligence in election campaigns. The ad, which portrays Tallarico rewriting biblical texts and drawing himself in crayon, is entirely AI-generated, raising concerns about the authenticity of political messaging. As political groups invest heavily in AI-generated content—over $100 million in the current midterm season—many of these ads do not disclose their artificial origins, complicating the landscape of political communication.
Research conducted by the Brennan Center for Justice reveals both the potential and pitfalls of AI in elections. While chatbots were tested for their ability to debunk election conspiracies, they often provided accurate information and pushed back against false claims. However, when tasked with creating disinformation, some chatbots, like Grok, complied and even enhanced the prompts with convincing details. This duality poses a significant challenge: while AI can assist in verifying information, it can also be weaponized to spread misinformation.
Experts warn that the proliferation of AI-generated content could undermine public trust in elections. Tools like TrueMedia are being developed to help identify manipulated media, but the gap between generative AI capabilities and detection tools remains concerning. As the 2024 elections approach, the emphasis on verifying information through official channels rather than relying solely on AI is crucial for maintaining electoral integrity. The ongoing evolution of AI in political contexts underscores the need for vigilance and critical engagement with the information presented to voters.
Show Notes
You used to need a government-backed operation to spread election disinformation. Now a chatbot might do. Larry Norden of the Brennan Center put the major AI chatbots to the test: Could they create election fakes? Debunk them? Even recognize their own handiwork? The answers were not all reassuring.
Learn about your ad choices: dovetail.prx.org/ad-choices
More from Click Here
What if Russia didn’t have to convince you of anything new? What if it could simply exploit what you already believe? That’s the idea behind a Russian-linked disinformation network that hijacks real social media conversations to spread propaganda and influence U.S. elections. We follow the trail from celebrity rumors to Congress, and ask whether AI can spot the manipulation before it spreads.
In Social Security’s database, a few keystrokes can turn a living person into a dead one. As part of a deportation scheme, DOGE pushed to do just that to 2.7 million people — without checking whether they were actually dead. Social Security whistleblower Jeremiah Schofield decided to stop them.
Deepfakes are getting harder to spot. But digital forensics expert Hany Farid says the bigger problem isn’t just that AI can fool us. It’s that even when something is real, people may refuse to believe it. What happens when seeing is no longer believing—and the truth has a hard time catching up?
In 30 years of representing whistleblowers, national security attorney Mark Zaid has never had so many government lawyers come to him for advice. They’ve seen things inside the Trump administration that trouble them. And they want to blow the whistle. But as lawyers, they know the law. Which, as it turns out, may be part of the problem.
Frances Haugen walked out of Facebook with more than 20,000 pages of internal documents. And somehow, nobody caught her. Now her story is getting the Hollywood treatment in Aaron Sorkin’s new movie, The Social Reckoning. But how did she actually pull it off? Haugen takes us behind the scenes of her surprisingly low-tech operation—and explains why the next Facebook whistleblower might not be so lucky.