Driverless cars hit a regulator gap

Politico

16 hours ago

For Silicon Valley, the promise of the second Trump administration is that it would cull all those pesky bureaucrats impeding innovation.
But sometimes, a new industry needs bureaucrats to cut red tape.
The autonomous vehicle industry could soon face this catch-22. As driverless cars continue to roll out in American cities, AV companies need the help of federal regulators to deal with the landscape of safety rules. Yet my colleague Pavan Acharya reported last week that the Transportation Department's Office of Automation Safety, which helps regulate AVs, is on the verge of becoming a ghost town.
The Republican-led Senate Appropriations Committee noted in a report that the office had 'lost almost all of its staff,' and encouraged the department to prioritize hiring.
Silicon Valley has generally pooh-poohed regulators as getting in the way of both moving fast and breaking things. However, it's regulators who are trying to work around policies designed for analog cars, so that AVs can start operating legally within the federal highway rulebook.
'[The Transportation Department] is taking away existing regulations that would stand in the way of autonomous vehicles,' said Ryan Calo, a University of Washington law professor who specializes in tech policy. 'They're trying to figure out how to revise their own code so that this is possible. Reducing personnel is just stupid.'
The Trump administration has been going all out on kneecapping the bureaucratic state, cutting tens of thousands of people from the federal workforce. Tesla CEO Elon Musk's DOGE is central to these efforts, and the tech industry has cheered it on.
In fact, Musk is apparently responsible for the exodus of AV regulators. The Financial Times reported in April that DOGE had dismissed about 30 people from DOT's National Highway Traffic Safety Administration, many of whom were employed at its Office of Automation Safety. NHTSA has been a cloud over Musk's dream of deploying fully self-driving Teslas, recalling 2 million of its vehicles in 2023 for issues with its autopilot feature. The FT reports that Musk felt regulators were being unfair, though there were also concerns within Tesla that gutting the automation safety office could backfire. (Tesla did not respond to DFD's inquiry.)
Those concerns may be merited. The Biden administration launched the safety office in 2023 specifically to help the AV industry overcome regulatory hurdles. It fields petitions from manufacturers for exemptions to the Federal Motor Vehicle Safety Standards, which sets performance and design requirements for cars operating in the U.S. The industry has long complained that the Transportation Department has been slow to process the petitions.
'It takes a lot of people to work through [exemption petitions] point-by-point,' said Philip Koopman, a Carnegie Mellon professor who has been the architect of industry safety standards for AVs. 'It's just labor intensive.' He cautioned, however, that the department could simply relax the exemption standards to the extent that you'd need just enough staff to press a rubber stamp.
Transportation Secretary Sean Duffy has already broadened exemptions to cover a wider range of AVs. The department just issued its first exemption under the expanded standards last week for Zoox, Amazon's robotaxi subsidiary – but the company is only allowed to test the vehicles. Getting an exemption to sell AVs or charge consumers for their services is a more rigorous process, though Duffy is working to streamline that too.
When DFD asked about the Senate appropriations report, NHTSA said in a statement that it has 'aligned resources accordingly, including staff, to add engineers to the Office of Automation Safety.'
Another reason why the industry might want the automation safety office to be staffed is that it could help clarify rules around AVs.
With the advent self-driving cars, companies face a conundrum when it comes to state versus federal laws. The current paradigm generally has the federal government setting nationwide standards for vehicles themselves. States then determine whether someone is qualified to operate those vehicles by issuing driver's licenses and operating permits. (My colleague Christine Mui recently reported on Tesla's troubles with California permitting.)
That regulatory divide gets fuzzier with AVs, especially for interstate travel, since the vehicle's AI becomes the driver. 'The design and construction of an autonomous vehicle, and driver competency is intertwined,' said Michael Brooks, executive director at the Center for Auto Safety. 'You can't really separate it into the traditional areas.' So when companies deploy a federally approved AV for a cross-country road trip, they don't necessarily know whether they're violating a patchwork of state laws.
Congress could resolve the ambiguity by passing legislation that empowers the Transportation Department to override state laws, according to Brooks, but there doesn't seem to be much progress on that front. In the meantime, having more federal automation safety staff could help the department proliferate nationwide regulations in a bid to assert its authority over the states. That would give AV companies a bit more legal certainty.
To be sure, staffing up the automation safety office wouldn't necessarily be a clear-cut benefit for the self-driving car industry. More regulators in general could mean more requests for data that companies would consider to be trade secrets, and more meticulous investigations of product defects.
But looking past the immediate policy barriers, having enough regulators to make sure that AVs are actually safe benefits all parties involved. 'If a driverless car is involved in an accident, people are going to get nervous about them,' said Calo. 'They need a trusted expert body that's impartial, like the Department of Transportation, to be able to look at those situations in a sober ride way and determine [...] what to do about it.'
STATE DEPARTMENT GETS INTO THE ONLINE SPEECH ARGUMENT
The GOP campaign against social-media moderation has trickled into formal American diplomacy: The Trump administration is now saying human rights in the United Kingdom 'worsened' in 2024 due to speech restrictions, particularly those online, POLITICO's Mizy Clifton reports.
The State Department on Wednesday released its annual assessment of countries' human rights practices in which it criticized the U.K. for chilling speech on social media surrounding incidents like the 2024 stabbing of three children in Southport last year. The report also targeted the country's Online Safety Act that rolled out last month, which requires platforms to verify users' ages in an attempt to prevent children from seeing violent or sexual content. The department suggested that the act could lead to 'government regulation to reduce or eliminate effective encryption (and therefore user privacy) on platforms.'
The safety act has already drawn criticism for age-gating certain content, such as posts on X about violence in Gaza or Reddit content on cigars. (Wikipedia also lost a case on Monday challenging the act's identity verification rules.) A congressional free speech delegation traveled to meet with U.K. officials about the act late last month.
Sen. Ted Cruz pushes for AI sandboxes
Senate Commerce Chair Ted Cruz wants to spur the development of artificial intelligence by exempting the tech from existing regulations, according to details of a draft bill shared with POLITICO's Benjamin Guggenheim and Anthony Adragna.
The bill would allow federal agencies to grant temporary waivers on their rules for companies to pilot AI systems. States have previously granted such exemptions for therapy chatbots and diagnostic tools that use the technology — an idea that tech companies love, but which critics say has some real flaws in practice.
Cruz prefers a light-touch approach to AI regulation. He was one of the main advocates for including a 10-year state AI law moratorium in the Big Beautiful Bill this summer, though the provision was dropped amid Republican infighting. Cruz's new legislation dovetails with President Donald Trump's call for regulatory sandboxes in his AI Action Plan to 'rapidly deploy and test AI tools.'
post of the day
THE FUTURE IN 5 LINKS
Stay in touch with the whole team: Aaron Mak ([e-mail]); Mohar Chatterjee ([e-mail]); Steve Heuser ([e-mail]); Nate Robson ([e-mail]); and Daniella Cheslow ([e-mail]).