A dire warning from Anthropic and OpenAI that increasingly autonomous AI agents could launch cyberattacks – or worse – has an especially chilling implication for Hollywood, where studios and streamers are rapidly weaving AI into their businesses. "Imagine a world in which there are thousands of Sony hacks happening in real time," Reel AI columnist Erik Barmack tells Elaine Low and Sean McNulty. "You could take down the entire industry. Confidentiality, protections, the secure release of IP – all of those things could be harmed."And while the guilds fight generative AI and Hollywood hashes out rules for how the technology can be used, Erik says the industry may be focused on the wrong threat: "What happens when recursive learning gets better and better, and these agents go off and do thinking on their own?"Plus: There may finally be momentum behind a federal film and TV tax incentive. Rep. Laura Friedman tells Elaine that lawmakers from both parties are working with Hollywood's guilds and Jon Voight – Donald Trump's special Hollywood ambassador, whom she describes as a key bridge to the president. Friedman says she hopes the proposal can move by year-end – and describes the unusual bipartisan effort as giving lawmakers "an extra sort of joy" in working together: "We're still one country."Learn more about your ad choices. Visit megaphone.fm/adchoices