State AI laws' authors ask frontier labs to slow down
The five state legislators who wrote America's toughest AI rules have a new message for frontier labs: pace yourselves. The Cybercab, meanwhile, gets a federal probe.
The architects of the country's strongest state AI laws are publicly asking frontier AI companies to slow down until their safety work can match the pace of their models. In an interview with NBC News, the five state legislators behind the most significant US state AI regulations said frontier AI companies need to collaborate on "pacing" the speed of AI development. The appeal comes as the same labs have spent the past two weeks shipping GPT-6 Astra, the latest Muse models from Meta, and a wave of new agentic features that have run ahead of any single federal rulebook. It is a striking inversion: the people who wrote the rules now want the companies to enforce them on themselves. With the federal moratorium stalled, a state-driven slowdown request is the closest thing to a regulatory speed bump the industry is likely to see this quarter.
NHTSA has opened a formal investigation into Tesla's Cybercab rollout in Austin. The agency said Friday it will examine "the process and technical data on which Tesla relied when certifying the Cybercab" — specifically whether Tesla's self-certification that the vehicle complies with all Federal Motor Vehicle Safety Standards was sound for a vehicle built without a steering wheel or pedals. Tesla told NHTSA it self-certified the Cybercab under FMVSS, the same route automakers normally use. The probe comes three days after Tesla began commercial rides with a 45-car fleet. Federal regulators have spent the past year carving out exceptions for AVs — the Trump administration proposed axing the brake-pedal requirement in June — but a launch with literally no driver controls is the first real test of where that lane ends.
Uber and Wayve beat Waymo to London with the UK's first commercial self-driving taxi service. A small fleet of Wayve-powered Ford Mustang Mach-Es started taking fares through UberX, Comfort and Electric on Wednesday, each with a safety operator behind the wheel. Wayve's AV2.0 system relies on cameras and radar instead of lidar and HD maps — an AI built to generalize to unmapped streets, not memorize them. Waymo, which is widely seen as the technical frontrunner in the US, has tested in London for months but hasn't started paid service. The launch matters because London is a regulatory environment few US robotaxi operators have cleared: complex junctions, weather, and streets never designed for the Google-style high-definition maps Waymo depends on.
What to watch: Whether the state-lawmaker appeal triggers any voluntary commitment from frontier labs before federal AI moratorium talks resume.
Will the Cybercab probe slow Tesla's rollout, or harden into the legal precedent that AVs need at least one human control?
