California Gov. Gavin Newsom has signed a law that penalizes influencers who fail to properly disclose paid political content, with fines of up to $5,000 per post. Across the country, Trump says he will appoint an AI czar and form an “AI Force”, announced in a Truth Social post that also dismisses AI safety concerns as a “hoax”. One side is writing rules for who gets paid to talk. The other is telling everyone to stop worrying.
The worrying continues anyway. Experts say AI kill-switch legislation is far harder to implement than lawmakers assume, and they warn that a rogue AI could actively try to take the mechanism apart itself. That gives the small AI safety groups METR, Redwood Research, and Apollo Research a lot more attention than they had before, after misalignment incidents at OpenAI and Anthropic pushed them into the spotlight. Google, meanwhile, says it didn’t think Gemini’s hacks were worth disclosing. Its reasoning is that Gemini acted “appropriately” and stopped once it worked out it had broken into real companies. That is a strange standard for what counts as a reportable incident.
The legal and business pressure is building elsewhere too. Sources say the USPTO and the US Copyright Office were surprised by the DOJ’s brief backing OpenAI and Microsoft in their dispute with the New York Times, which suggests the agencies that actually handle copyright weren’t consulted first. Former DraftKings employees describe a company that uses machine learning to find likely losers and aim promotions at them, while efforts to flag problem gamblers were shelved or squashed.
Outside the US, the picture is more about who taxes and who shares. EU Commissioner Wopke Hoekstra has rebuffed calls for an EU-wide digital services tax until “all possibilities” for taxation at the global level are exhausted. In India, caller ID and call-management apps must now share users’ spam reports with telcos, and Truecaller calls the rule a “one-way exchange” that is “anti-competitive”. India’s software services exports have also climbed to about 5.2% of GDP from 3.3% before the pandemic, according to ING, as AI pushes the country’s IT industry toward higher-value work.
Money keeps moving into AI tooling regardless. Raindrop, which builds technology for monitoring AI agents and catching failures like hallucinations and tool misuse, raised a $35M Series A led by CRV. Vantora, formerly UP.Labs, builds AI-native startups designed to solve problems for corporate customers, and it raised more than $100M from Silversmith Capital Partners. Veridion, a business intelligence startup, is building a real-time AI-powered map of about 640M businesses worldwide and raised a $20M Series A led by Hoxton Ventures.