Online Safety Bill doesn't make any sense, says former Facebook security officer Alex Stamos

Summary

Alex Stamos, former Facebook chief security officer, criticizes the U.S. Online Safety Bill as contradictory and unlikely to pass the Senate, and blasts the White House's handling of AI regulation. He argues the administration's abrupt ban on an Anthropic model created political risk for American AI companies while Chinese AI models have rapidly closed the gap, all under a regulatory vacuum.

• The Online Safety Bill is a mashed-together collection of bills that requires protecting kids but forbids ID verification, making it unworkable.
• Stamos believes the bill is merely a messaging exercise and will not pass the Senate, minimizing new regulatory risk for social media.
• The White House's last-minute ban on Anthropic's Fable model, based on Amazon complaints, was an overreaction that created huge political uncertainty for U.S. AI developers.
• Two Chinese AI models have since been released with capabilities, including cyber capabilities, matching or exceeding the best American models.
• The administration scrapped the prior AI rulemaking process, fired relevant staff, and has issued no clear rules for what American model makers can provide.
• Stamos says the U.S. is allowing China to buy its best chips while stifling domestic AI, calling the overall policy contradictory.