3-Shot Learning #6
Deepfake legislation, Taylor Swift scams, and US-China diplomacy
Welcome to 3-Shot Learning, a weekly newsletter from the Center for AI Policy. Each issue explores three important developments in AI, curated specifically for AI policy professionals.
Morelle and Kean Push Congress to Legislate Against Deepfakes
Back in December of 2022, Representative Joe Morelle (D-NY) introduced the “Preventing Deepfakes of Intimate Images Act,” a bill that would outlaw the nonconsensual circulation of sexual content produced using digital manipulation, and establish legal avenues for victims to seek justice. The original bill garnered zero cosponsors (which is not exactly shocking, given that Morelle introduced it just two weeks before the end of the Congress). Since then, Morelle has not only reintroduced the bill, but has also attracted media coverage in major outlets and gained Republican support from Representative Tom Kean, Jr. (R-NJ), who hosted a joint press conference with Morelle on Tuesday to discuss the bill.
This is not Kean’s only effort to regulate AI-generated content, since he has also introduced the “AI Labeling Act.” The bill, true to its name, would require generative AI outputs to include conspicuous disclosures that clearly label the content as AI-generated. Additionally, the bill would establish a NIST-led working group to produce guidance for online platforms in detecting AI-generated content. Notably, Senators Brian Schatz (D-HI) and John Kennedy (R-LA) have introduced identical legislation in the Senate, with support from influential groups like the AFL-CIO.
Kean’s interest in the topic likely intensified after deepfakes caused distress in his district. Last fall, the town of Westfield, New Jersey made national news after a student used AI to create and distribute nude photographs of female classmates. One of the targeted girls, 14-year-old Francesca Mani, responded to the event by contacting area lawmakers and urging them to take action. Francesca’s efforts caught the attention of Representatives Kean and Morelle, who included her in their Tuesday press conference. But it remains to be seen whether America’s full legislative branch will work together and actually pass legislation on this issue.
Taylor Swift Le Creuset Scam Thrives on Meta
Since last month, scammers have been using AI-generated audio to mimic Taylor Swift’s voice within video advertisements for a (fake) giveaway of Le Creuset cookware. The videos lure victims into paying a small one-time “shipping fee” that automatically converts into an expensive monthly subscription. Unfortunately, this certainly won’t be the last time that AI-empowered scammers swindle Americans—just last week, 404 Media reported how AI clones of Taylor Swift (again), Joe Rogan, and others are “hawking Medicare and Medicaid scams to millions of people on YouTube.” Overall, AI tools seem poised to play a growing role in fraud, which the FTC reports cost $8.8 billion in 2022 alone.
US and Chinese AI Experts Have Been Meeting to Discuss AI Risks
The Financial Times described last week how “scientists and policy experts” from leading AI companies OpenAI, Anthropic, and Cohere met twice last year with Chinese experts from Tsinghua University and other state-sponsored Chinese groups. Certain government officials in the US, China, and the UK were fully aware of these meetings, and therefore presumably approved of this diplomatic activity. Despite contentious military and trade relations between the US and China, the discussions—which focused on AI risks and AI safety research—were successful enough that the participating groups are planning to do more meetings in the future. These ongoing meetings could play an important role in broader multi-track diplomacy on AI between the US and China, where it remains unclear what agreements the two superpowers can reach in the next few years.
News at CAIP
We had a great turnout at our happy hour last week, which we hosted at Sonoma Restaurant & Wine Bar.
We commented on the Commerce Department’s revised export controls affecting AI hardware. Read our full comments here.
Quote of the Week
You have on the one hand all of the AI cheerleaders who say that “AI is going to be wonderful, AI is going to be life-changing; it’s going to be world-changing, it’s going to be the best thing that has ever happened to the human race” or something to that effect. [...] Then when you leave the confines of this building and the lobbyists who inhabit it—when you go out and actually talk to real people working real jobs—you find the second set of concerns. I have yet to talk to a Missourian who is an enthusiastic, no-holds-barred cheerleader for AI.
—Senator Josh Hawley (R-MO) in last week’s hearing on AI and the future of journalism
This edition was authored by Jakub Kraus.
If you have feedback to share, a story to suggest, or wish to share music recommendations, please drop me a note at jakub@aipolicy.us.
—Jakub