States Steer Autonomous Vehicle Legislation
Key highlights this week:
We’re currently tracking 692 bills in 45 states related to AI this year, 111 of which have been enacted into law.
The California CPPA will address automated decision-making technology at an upcoming hearing.
Delaware governor signs political and sexual deepfake bills into law and lawmakers send a sexual deepfake bill to the governor in Pennsyvalnia.
Upcoming study committee meetings addressing AI in Georgia and New Jersey.
While much of the focus on AI this year has been on the newest trend of generative AI, states have also addressed an AI-related legislative trend with more history — autonomous vehicles (AVs). As manufacturers continue testing and deploying AVs on public roads, most states now have laws in place regulating the operation of AVs on their public roads.
To start, let’s briefly review how various AVs (or “robotaxis” or “self-driving cars” as they’re often referred to in the press) are classified. AVs are classified as one of six levels established by SAE International. Levels 0-2 are cars that require a human driver to perform all of the driving functions at all times but may have some autonomous features such as lane departure warnings or autonomous parallel parking. Level 3 cars can operate autonomously in certain situations and environments and require a human driver to be able to take over driving functions at any moment. Finally, level 4-5 cars are capable of performing all driving tasks in most (level 4) or all (level 5) environments and weather conditions. Currently, there are no level 4 or 5 AVs for sale to private individuals and many of the AVs being tested are classified as level 3 autonomous vehicles with safety drivers behind the wheel. Waymo and Cruise are the only two major AV companies that have deployed level-4 AVs available to the public in select cities.
While Congress has considered autonomous vehicle legislation in the past, states have led the way in enacting AV legislation. Currently, 27 states and the District of Columbia allow the testing or operation of level 4 and 5 vehicles on public roads, while an additional 6 states allow the testing or operation of only level 3 AVs.
This year, South Dakota (SD HB 1095), Kentucky (KY HB 7), and Alabama (AL SB 226) enacted legislation that allows level 3-5 AVs to be tested and operated on public roads. Each of these laws requires that AVs operating on public roads can comply with all traffic and motor vehicle safety laws, specifies how an AV is to respond if it is involved in an accident, and establishes insurance requirements. Additionally, the Kentucky law requires AV manufacturers to develop a law enforcement interaction plan that describes how law enforcement will communicate with an AV fleet support specialist and how to safely remove an AV from the road if necessary.
While most states were able to pass AV legislation with relatively little pushback, lawmakers enacted the Kentucky bill after overriding Governor Andy Beshear's (D) veto. While the bill is very similar to the more than 30 AV laws currently in place throughout the country, Governor Beshear stated in his veto message that the bill did not address questions concerning the safety and security of AVs. The legislature disagreed with his assessment and ultimately overrode his veto, which only requires a simple majority of legislative votes in Kentucky.
California enacted AV legislation in 2012 (CA SB 1298) and currently has several companies testing and operating AVs on public roads. This year, the legislature passed four AV bills, two of which Governor Gavin Newsom (D) signed into law (CA SB 1313 and CA AB 1777), and two were vetoed by the governor (CA AB 2286 and CA AB 3061). The first bill (CA SB 1313) signed into law prohibits buying or selling a device that is designed to neutralize or disable driver assistance technology or autonomous vehicle technology. The second bill that became law this year (CA AB 1777) authorizes peace officers to issue notices of noncompliance if they observe a vehicle that has violated state or local traffic laws. This bill is particularly noteworthy as residents of San Francisco have complained in recent months about AVs being tested on city streets getting stuck or behaving erratically.
As we briefly mentioned in our last update, Governor Newsom vetoed two AV-related bills. One (CA AB 3061), which would have established reporting requirements for AV collisions and disengagements, and another bill (CA AB 2286) that would have prohibited AVs weighing more than 10,001 pounds and used to transport goods or passengers from being operated or tested on public roads without a safety driver. The governor vetoed a very similar bill last year, which also would have restricted the use of fully autonomous trucks in the state.
Finally, the District of Columbia, which first enacted AV legislation in 2013 (DC B 19-278), enacted temporary legislation (DC B 709) requiring safety drivers to be physically present in an AV and requires the District Department of Transportation (DDOT) to be notified at least ten business days before testing or operating an AV. This emergency legislation is currently in force until January 12, 2025, and addresses concerns raised by DC City Council members regarding AV testing occurring before AV testing rules are established by DDOT. The DC City Council is currently considering legislation to make this requirement permanent (DC B 710) and while that bill has not received much attention since having a hearing in June, it is likely that the City Council will move the legislation forward during the lame duck period after the election next month.
As we head into 2025, AV companies continue to expand their testing and deployment of this technology on public roads, which should lead to more lawmakers taking up the issue either to enact new testing and operation requirements or to amend current requirements as these vehicles interact with the public.
Recent Developments
In the News
Comments Criticize FCC Proposed Rule on Political Deepfakes: Initial comments on a September proposal by the Federal Communications Commission (FCC) to require callers to disclose to consumers when they receive AI-generated robocalls or text messages were due on Oct. 10. Groups were both supportive and critical of the proposed rules, some supporting the goals of the proposal but calling the specifics overly broad. Reply comments are due by Oct. 25.
Major Policy Action
California: The California Privacy Protection Agency (CPPA) will have a November 8 hearing to discuss proposed regulations on automated decision-making technology (ADMT) that would impose audit and risk assessment obligations, and require notice to the consumer with an opportunity to opt out of the use of ADMT. The regulations were originally proposed in 2023 and were advanced earlier this year after an estimated 3,000 pages of comments from stakeholders.
Delaware: Gov. John Carney (D) signed two deepfakes bills into law last week that would take effect immediately. One bill (DE HB 316) addresses political deepfakes, requiring a disclaimer for communications created or manipulated within 90 days of an election. Another bill (DE HB 353) creates civil and criminal remedies for the wrongful disclosure of sexual deepfakes.
Pennsylvania: Lawmakers sent a sexual deepfake bill (PA SB 1213) to the governor for his signature to become law. The bill would make it a crime to disseminate, intentionally view, or possess artificially generated sexual depictions of an individual and would add artificially generated depictions to provisions prohibiting child pornography.
Georgia: The Senate Study Committee on Artificial Intelligence will hold its next hearing on Wednesday, October 23. The panel met earlier this month to learn about AI use in education and regulatory frameworks from around the world.
Notable Proposals
New Jersey: The Assembly Science, Innovation and Technology Committee will hold a hearing on Monday, October 21, to consider artificial intelligence legislation. One bill on the agenda (NJ AB 4639) would require boards of education to adopt policies on the use of video surveillance systems with artificial intelligence capabilities, while a resolution (NJ AR 158) would urge generative artificial intelligence companies to make voluntary commitments regarding employee whistleblower protections.