NHTSA Opens New Probe into Tesla’s Autopilot Following More than a Dozen Fatal Accidents

The NHTSA has linked Tesla's Autopilot to over a dozen fatalities and hundreds of crashes, prompting a new investigation into the adequacy of Tesla's December recall of 2 million vehicles. The probe reports that Tesla’s driver-assist system led to avoidable crashes involving visible hazards, suggesting a critical safety gap between driver expectations and the system’s capabilities. The investigation will assess if Tesla’s recall remedies were sufficient to address these safety risks.

Risk classification

• Primary risk domain: 5 Human-Computer Interaction
• Primary risk subdomain: 5.1 Overreliance and unsafe use

The core issue identified by regulators is a critical safety gap where drivers overly trust and rely on the Autopilot system, leading to inattention and foreseeable misuse of a system that is not fully autonomous.

Additional risk subdomains

• 7.3 Lack of capability or robustness: The Autopilot system exhibited a lack of capability to reliably detect and avoid visible hazards, such as stationary emergency vehicles, leading to avoidable crashes.

Causal factors

• Entity: AI
• Intent: Unintentional
• Timing: Post-deployment

The safety risks and crashes occurred after the Autopilot system was deployed on public roads, resulting from unintended outcomes of the system's operation and driver interaction.

EU AI Act risk tier

• Risk tier: 2 High Risk

High Risk: The report describes an AI system used in critical transport infrastructure and vehicle safety, which has significant implications for physical safety and public interest, and is subject to stringent regulatory oversight.

AI system and alleged parties

• AI system: Autopilot (Tesla)
• AI purpose: Autonomous Driving; Navigation Assistant
• Behaviour type: Assistant
• Alleged developer: Tesla
• Alleged deployer: Tesla drivers, Tesla
• Alleged harmed parties: Tesla drivers, General public, Drivers

Harm severity

Severity ratings are based on this scale.

Highest direct severity in any category: Substantial. Severity is scored from Negligible to Catastrophic in each harm category, for harm the reports describe as caused directly or indirectly by the AI system.

• Physical: direct Substantial, indirect Negligible
• Infrastructure: direct Negligible, indirect Negligible
• Property: direct Substantial, indirect Negligible
• Financial: direct Negligible, indirect Negligible
• Environmental: direct Negligible, indirect Negligible
• Malicious content: direct Negligible, indirect Negligible
• Differential treatment: direct Negligible, indirect Negligible
• Civil rights: direct Negligible, indirect Negligible
• Democracy: direct Negligible, indirect Negligible
• Privacy: direct Negligible, indirect Negligible
• Psychological: direct Negligible, indirect Negligible
• Epistemic: direct Negligible, indirect Negligible
• Child sexual exploitation and abuse: direct Negligible, indirect Negligible

Physical

Reported: The report explicitly describes physical harm, linking the Autopilot technology to at least 14 fatalities, several dozen injuries, and hundreds of crashes.

Directly caused: At least 14 fatalities and several dozen injuries resulted from crashes where Autopilot was engaged, including a fatal 2018 crash involving an Apple engineer.

Inferred additional harm: It is highly likely that additional injuries and fatalities occurred across the 467 identified crashes that were not individually detailed in the reports.

Property

Reported: The report explicitly describes property damage resulting from hundreds of crashes.

Directly caused: Hundreds of Tesla vehicles and other involved vehicles sustained severe damage or were destroyed in the 467 identified crashes.

Inferred additional harm: Significant property damage totaling millions of dollars is inferred across the hundreds of crashes involving high-value electric vehicles.

People affected

• Occurrences reported: 467
• People reportedly harmed: 50
• People reportedly exposed: 2000000

Potential causes

Management

• Misleading Marketing Claims: Management overstated system capabilities, creating false expectations.
• Disagreement with Safety Analysis: Management resisted regulator findings, slowing down effective remedies.

Technology

• Weak Driver Engagement System: System failed to effectively monitor and ensure continuous driver attention.
• Permissive Operating Capabilities: Autopilot allowed execution of tasks beyond its safe operational limits.
• Inadequate Recall Software Update: Software update failed to resolve the core safety and engagement issues.

Human Factors

• Driver Inattentiveness: Drivers failed to maintain continuous active supervision of the vehicle.
• Over-reliance on Automation: Drivers overly trusted the system due to misleading naming and marketing.
• Foreseeable Driver Misuse: Drivers misused the system by treating it as fully autonomous.

Process and Methods

• Reversible Safety Controls: Recall remedies allowed drivers to easily reverse critical safety updates.
• Opt-in Safety Remedies: Key portions of the safety recall required manual owner opt-in.
• Inadequate Recall Testing: Preliminary agency testing showed the recall did not fix safety risks.

Regulatory Environment

• Delayed Regulatory Action: Regulators investigated for three years before enforcing a recall.
• Lack of Standardized Names: No regulations prevented misleading names like Autopilot or FSD.

Information quality

• Classification confidence: High
• Reason for confidence: The reports from reputable news organizations provide detailed, consistent information regarding the NHTSA investigations, specific crash statistics, fatalities, and the technical nature of the safety gap identified by regulators.
• Ambiguities identified: There is minor variation in the reported number of fatalities (13 vs 14) depending on the specific investigation window referenced.
• Alternative interpretations: None. The incident is clearly a human-computer interaction and system safety failure.

The NHTSA investigation into Tesla's Autopilot recall highlights safety gaps in consumer AI driver-assist systems. While causing physical harm and property damage, the incident is a consumer safety and regulatory matter with negligible direct implications for national security.

• Overall national security impact: Minor
• Scope: Single nation
• Primary target: United States
• Alleged perpetrator: Unknown

Threat characteristics

• Imminence: Long-term. Represents an ongoing safety and regulatory concern rather than an active national security crisis.
• Autonomy: Human-supervised. The system operates with autonomous steering and acceleration but is designed to require active human driver supervision and intervention.
• Novelty: Established threat. Driver-assist safety failures and subsequent regulatory investigations are well-established issues with existing precedents.

Impact by dimension

• Physical security: Minor. The incident involves physical crashes and fatalities on public roads, but it remains a consumer safety issue managed under standard regulatory procedures rather than a national security threat to critical infrastructure.
• Information security: Negligible. No implications for intelligence capabilities, classified information, or information warfare operations.
• Sovereignty: Negligible. No threat to state authority, territorial control, or core government operations.
• Economic security: Minor. Represents financial and regulatory risk for a major domestic automotive manufacturer, but does not threaten overall national economic stability or strategic technological advantage.
• Societal stability: Negligible. Public safety concerns are present, but there is no threat to social cohesion, civil liberties, or large-scale human rights.