Two senior U.S. senators have formally requested that the National Highway Traffic Safety Administration (NHTSA) open a detailed investigation into Tesla’s public safety claims regarding its Full Self-Driving (FSD) technology, specifically targeting the automaker’s methodology for computing crash rates. Massachusetts Senator Ed Markey and Connecticut Senator Richard Blumenthal dispatched a letter to NHTSA Acting Administrator Ann Carlson on June 16, 2026, urging the agency to scrutinize what they describe as potentially deceptive comparisons between Tesla vehicles operating under automated driving modes and the general U.S. vehicle fleet.

The senators’ action follows months of growing skepticism from safety advocates and independent researchers who argue that Tesla’s “crash math” paints an overly rosy picture of its autonomous systems by comparing highway miles—where Autopilot and FSD are predominantly engaged—against all driving environments, including city streets, rural roads, and adverse weather. Markey and Blumenthal are calling on NHTSA to evaluate whether such statistical framing misleads consumers and regulators, and to determine if Tesla is in violation of federal safety disclosure obligations.

Background on Tesla’s FSD Safety Assertions

For years, Tesla has published quarterly safety reports that indicate vehicles using Autopilot or FSD Beta experience significantly fewer crashes per mile than the average vehicle in the United States. The company’s data, typically presented as accidents per million miles driven, often shows a roughly fivefold improvement over human drivers. Elon Musk and Tesla executives regularly cite these figures during earnings calls, product launches, and public appearances to bolster the case for autonomous driving’s life-saving potential.

However, critics have pointed out that Tesla’s comparisons are statistically flawed. The vast majority of Autopilot-enabled miles are logged on limited-access highways—environments inherently safer than the average road in America. By contrasting highway-heavy Tesla data with a national crash rate that includes all road types, Tesla’s methodology inflates the apparent safety margin. Furthermore, the company does not publish raw data or detailed breakdowns of disengagement events, near-misses, or crashes that are less severe but still reportable under NHTSA’s Standing General Order on automated-vehicle crashes.

The Senators’ Key Demands

In their letter, obtained by windowsnews.ai, Markey and Blumenthal lay out a series of requests. They ask NHTSA to:

  • Audit Tesla’s crash-rate calculations for both Autopilot and FSD Beta, including the underlying data filtering and normalization methods.
  • Determine whether Tesla’s public communications constitute false or misleading advertising under federal consumer protection laws.
  • Require Tesla to submit raw telemetry data from all FSD-equipped vehicles involved in collisions, including instances where automated systems disengaged just before impact.
  • Assess whether Tesla’s “Safety Score” driver-rating system—used to determine eligibility for FSD Beta—is scientifically valid and transparent.
  • Investigate reports of crashes where FSD Beta was active but not reported to NHTSA in a timely manner, as mandated by the 2021 automated vehicle crash reporting order.

The senators also emphasized the need for NHTSA to probe the “crash math” itself: the possibility that Tesla selectively excludes certain types of incidents or fails to account for increases in rear-end or stationary-object collisions that are overrepresented in partial automation systems.

“Tesla’s persistent overstatement of its vehicles’ autonomous capabilities, backed by unverifiable statistics, poses a direct threat to public safety,” the letter states. “Americans deserve to know whether these technologies are genuinely safer, or whether they are being marketed on a false premise.”

The Crash Math Controversy Explained

The dispute over Tesla’s crash-rate comparisons has simmered for the better part of a decade. Independent analyses from organizations such as the Insurance Institute for Highway Safety (IIHS) and AAA have repeatedly warned that partial automation can create a false sense of security, increasing driver distraction. A 2024 study from the Virginia Tech Transportation Institute, funded by NHTSA, found that after adjusting for road type, driver age, and vehicle class, the safety gains of Level 2 systems like Tesla’s Autopilot were statistically negligible.

Tesla’s methodology faces even sharper critique when applied to FSD Beta—a feature marketed as capable of navigating complex urban environments with human-like judgment. Safety experts note that FSD Beta users are often early adopters who are more attentive and technically savvy than the average driver, meaning that any observed crash reduction may be due to driver demographics rather than system performance. Moreover, FSD Beta operates primarily on pre-mapped roads with clear lane markings, further skewing the data set toward safer road contexts.

“Tesla is comparing apples to apple-flavored candy,” says Dr. Liza Dixon, a human-factors researcher at the University of Michigan Transportation Research Institute, who has not seen the senators’ letter but has publicly criticized the automaker’s safety claims. “Until we have access to the full denominator—disengagements, near-misses, and actual crash causes—these numbers are marketing, not science.”

NHTSA’s Ongoing Scrutiny of Tesla

NHTSA has multiple open investigations into Tesla’s driver-assistance systems. The agency’s defect probe into Autopilot, launched in 2022 and later expanded, has examined dozens of crashes involving Tesla vehicles striking stationary emergency vehicles. In 2024, NHTSA upgraded that investigation to an engineering analysis, a step just short of a formal recall demand. The agency has also demanded detailed data from Tesla as part of its broader inquiry into all Level 2 systems across the industry.

Separately, NHTSA’s Special Crash Investigations unit has examined over 40 Tesla crashes where advanced driver-assistance systems were believed to be engaged at the time of the collision. The agency’s published ODAP (Office of Defects Analysis and Prevention) reports have occasionally noted that in some incidents, Tesla failed to provide crash notification within the required 24-hour window.

The senators’ new letter could accelerate these probes or prompt a new, focused investigation into FSD Beta specifically. It also adds political weight: Markey and Blumenthal sit on the Senate Commerce, Science, and Transportation Committee, which oversees NHTSA and has held hearings on autonomous vehicle regulation.

Industry and Regulatory Context

The push for greater transparency around automated-vehicle safety metrics is not limited to Tesla. A proposed rulemaking from NHTSA, expected in late 2026, may mandate that all automakers report standardized safety data for systems above SAE Level 2, including miles driven in each operational design domain, disengagement reasons, and crash root causes when automation was active or recently disengaged.

Safety advocates have long argued that the current self-reported crash order—enacted in 2021—is insufficient because it allows manufacturers to define their own reporting criteria. “We need a level playing field where consumers can actually compare apples to apples,” said Cathy Chase, president of the Advocates for Highway and Auto Safety. “The senators are right to demand NHTSA dig deeper into Tesla, but this is an industry-wide problem.”

Separately, the Federal Trade Commission (FTC) has been investigating Tesla’s naming and marketing of “Full Self-Driving” since 2023, questioning whether the term constitutes a deceptive practice given that the SAE Level 2 system requires constant driver supervision. No final action has been taken, but the agency has issued civil investigative demands to Tesla regarding its advertising.

Tesla’s Likely Response

Tesla has historically defended its safety reports as accurate and transparent. In past communications, the company has argued that its crash-rate data is corroborated by third-party telematics studies and that critics focus on hypothetical risks rather than the billions of real-world miles demonstrating system capabilities. Tesla’s November 2025 Impact Report highlighted that vehicles with FSD Beta engaged experienced 5.2 times fewer crashes per mile than the national average—a metric the company says includes all crashes reported to its servers, regardless of severity.

In response to earlier NHTSA inquiries, Tesla has contended that its Autopilot disengagement data is proprietary and that releasing granular details could compromise its competitive advantage. The company also maintains that the safety of its system is best judged by the steep decline in crash rates over successive software updates.

Industry analysts, however, view the senators’ pressure as a significant escalation. “When members of Congress get involved, it often accelerates the regulatory timeline,” says Michael Ramsey, an autonomous-vehicle analyst at Gartner. “Tesla may find itself forced to open its black box sooner than it would like.”

What Comes Next

NHTSA has not yet formally responded to the senators’ letter. The agency typically acknowledges such congressional inquiries within 30 days and may incorporate them into existing investigations. Given the ongoing engineering analysis into Autopilot, it is possible that NHTSA will expand that probe rather than open a new one. But the explicit focus on “crash math” could lead to a separate, statistically-focused audit.

If NHTSA finds evidence of misleading safety claims, the agency could issue a consent order requiring Tesla to modify its public reports and pay fines. In an extreme scenario, NHTSA could refer the matter to the Department of Justice for false statements. The FTC, concurrently, might issue its own ruling on the FSD moniker. Both agencies have shown increased willingness under the current administration to pursue enforcement actions against tech-forward automakers.

For consumers, the inquiry raises immediate questions about whether FSD—which currently costs $12,000 as an upfront purchase or $199 per month—delivers on its safety promises. Early adopters who have long believed their vehicles are significantly safer may find that the statistical advantage is far narrower than Tesla claims.

Conclusion

The clash between Tesla’s narrative of autonomous safety and the demands of federal regulators and lawmakers is coming to a head. Senator Markey and Senator Blumenthal have shone a bright light on the methodological gaps in how crash statistics are framed. As NHTSA weighs its next steps, the outcome could redefine not only how Tesla markets its most advanced software but also the standards by which all automated driving systems are judged.

In an industry racing toward higher levels of automation, transparency around safety claims is essential. The senators’ intervention ensures that regulators will now dig into the math behind the marketing—and the results could change the conversation about self-driving cars for years to come.