This episode of the Cyberlaw Podcast kicks off with a spirited debate over AI regulation. Mark MacCarthy dismisses AI researchers' recent call for attention to the existential risks posed by AI; he thinks it's a sci-fi distraction from the real issues that need regulation – copyright, privacy, fraud, and competition. I'm utterly flummoxed by the determination on the left to insist that existential threats are not worth discussing, at least while other, more immediate regulatory proposals have not been addressed. Mark and I cross swords about whether anything on his list really needs new, AI-specific regulation when Big Content is already pursuing copyright claims in court, the FTC is already primed to look at AI-enabled fraud and monopolization, and privacy harms are still speculative. Paul Rosenzweig reminds us that we are apparently recapitulating a debate being held behind closed doors in the Biden administration. Paul also points to potentially promising research from OpenAI on reducing AI hallucination.
Gus Hurwitz breaks down the week in FTC news: Amazon has settled an FTC claim over children's privacy and another over security failings at Amazon's Ring doorbell operation. The bigger story is the FTC's effort to impose a commercial death sentence on Meta's line of children's advertising and services—for a crime that looks to Gus and me like a misdemeanor. Meta thinks, with some justice, that the FTC is just looking for an excuse to rewrite the 2019 consent decree, something Meta says only a court can do.
Paul flags a batch of China stories:
