AV Safety Data Act
- Sponsor
- Sen. Markey, Edward J. [D-MA]
- Committees
- Commerce, Science, and Transportation Committee (primary)
- Last action
- Jan 29, 2026
Bottom line
The AV Safety Data Act aims to enhance public safety and transparency by requiring more comprehensive and granular incident and operational data from autonomous vehicle companies, making this data publicly accessible for oversight and analysis.
What it actually does
This bill mandates the National Highway Traffic Safety Administration (NHTSA) to promulgate regulations requiring manufacturers and operators of autonomous vehicles (ADS and Level 2 ADAS) to submit monthly reports. These reports must include detailed mileage data, information on collisions involving vulnerable road users or other vehicle occupants, and comprehensive data on 'unplanned stoppage events.' The bill also requires NHTSA to make this collected data publicly available in a machine-readable format.
Proponents argue
Proponents argue this bill is crucial for public safety, transparency, and informed regulatory oversight of rapidly developing autonomous vehicle technology. They contend that current reporting is insufficient to identify safety trends, assess risks, and build public trust, especially concerning incidents not involving collisions but still posing significant hazards.
Opponents contend
Opponents, likely from the autonomous vehicle industry, might argue that the reporting requirements are overly burdensome, could reveal proprietary information, or that existing reporting mechanisms are sufficient. They might also express concerns about the potential for misinterpretation of raw data by the public or competitors, which could unfairly damage public perception.
The bill is concise and can be read quickly, allowing ample time for stakeholders to understand its provisions before implementation.
Section 2(a)
Definitions for Autonomous Vehicle Reporting
This section defines key terms used throughout the bill, such as 'Administrator' (NHTSA Administrator), 'covered entity' (manufacturers/operators subject to NHTSA's Third Amended Standing General Order 2021-01), and 'covered vehicle' (vehicles with Automated Driving Systems or Level 2 Advanced Driver Assistance Systems). Crucially, it also defines 'unplanned stoppage event' to include non-collision incidents where an AV stops in a travel lane or interferes with public services.
Supporters argue
Supporters argue clear and comprehensive definitions are essential for consistent and effective implementation of reporting requirements, ensuring all relevant parties understand their obligations and that critical safety events are captured.
Critics contend
No specific opposition is typically raised against definitions themselves, but the breadth of 'covered entity' and 'covered vehicle' (especially Level 2 ADAS) and the definition of 'unplanned stoppage event' are points of contention for industry regarding the scope of reporting.
Tradeoffs
Balancing the need for broad applicability to ensure comprehensive safety oversight with potential concerns about overreach into less autonomous systems or overly broad incident definitions.
Section 2(b)
Mandated Monthly Incident and Operational Data Reporting
This provision mandates NHTSA to issue regulations within 90 days requiring covered entities to submit monthly reports. These reports must include total miles traveled by covered vehicles (disaggregated by make, model, software version, road type, location, and occupant presence), information on collisions involving vulnerable road users or occupants of other vehicles, and detailed information on 'unplanned stoppage events.'
Section 2(c)
Limitations on Level 2 Advanced Driver Assistance System Data Reporting
This section specifies that data from Level 2 Advanced Driver Assistance Systems (ADAS) can only be submitted if it was collected while the system was engaged or during the 30-second period preceding an unplanned stoppage event. Crucially, it explicitly prohibits the submission of personally identifiable information about a human driver.
Section 2(d)
Public Access to Autonomous Vehicle Safety Data
This provision mandates that, starting 120 days after enactment, NHTSA must make all submitted information and reports publicly available on its website. The data must be provided in a machine-readable format and include datasets, facilitating analysis by researchers, safety advocates, and the general public.
Section 2(e)
Future Revision of Reporting Regulations
This section allows the Administrator, beginning 10 years after the date of enactment, to rescind or revise the reporting regulations to reduce the frequency and scope of information required. It also clarifies that NHTSA can revise regulations at any time in a manner consistent with the bill's overall intent, providing flexibility for future adjustments.
Defines 'unplanned stoppage event' to include non-collision incidents where a covered vehicle stops on a public road in a travel lane and is unable to proceed, or interferes with public transit, law enforcement, first responders, or construction zones, if the Automated Driving System or Level 2 Advanced Driver Assistance System was engaged or attempted to be engaged 30 seconds prior.
Section 2(a)(6)
Why it matters:While defining terms is standard legislative practice, the comprehensive and expansive nature of this particular definition is highly consequential for the scope of required reporting.
Case for: Proponents argue this definition is crucial for capturing a comprehensive picture of AV safety, as non-collision incidents can still create significant hazards, traffic disruptions, and risks to public safety, which traditional collision reporting would miss.
Case against: Industry might argue that this definition is overly broad, capturing minor inconveniences rather than true safety incidents, and could lead to an inflated perception of AV failures or an overwhelming volume of data that is difficult to analyze effectively.
Estimated impact: Significantly expands the scope of reportable incidents beyond traditional collisions, potentially leading to a much larger dataset of operational challenges and near-miss scenarios, providing a more holistic view of AV performance.