Robotaxi Safety Trackers and Robot Arrests: How We Monitor and Respond to Robot Incidents
As robots and autonomous vehicles become more common in public spaces, the question of how we track their safety record — and what happens when things go wrong — is becoming increasingly important. Two emerging developments illustrate how far we still have to go: independent safety trackers that monitor robotaxi crash data, and real-world incidents where police have physically intervened with robots.
Robotaxi Safety Trackers
When Tesla launched its robotaxi service in Austin, Texas in June 2025, independent researchers began tracking every reported incident through NHTSA filings. The resulting Robotaxi Safety Tracker has become a valuable resource for understanding how autonomous ride-hailing services perform in the real world.
As of mid-March 2026, the tracker has documented 15 incidents involving Tesla robotaxis in Austin. These include collisions with a bus, a heavy truck, a pole, and multiple fixed objects. Eight of the 15 crashes occurred at speeds below six miles per hour, and five at speeds between zero and two miles per hour. One incident was quietly upgraded to include a hospitalization injury without a public disclosure from Tesla.
The data raises questions. Estimated crash rates for Tesla’s robotaxi service come to roughly one crash per 57,000 miles — compared to approximately one crash per 500,000 miles for the average human driver. That is about nine times the human average. By contrast, Waymo — which operates a fully driverless fleet of over 2,500 vehicles and has logged more than 127 million autonomous miles — has demonstrated crash rates for injury-causing incidents that are 85 percent lower than human drivers, according to peer-reviewed research.
An important caveat: every Tesla robotaxi currently operates with a trained safety supervisor in the vehicle. Tesla announced unsupervised operation in January 2026, but as of early 2026 this applied to only a single vehicle in a limited area. Tesla also redacts all crash narratives filed with NHTSA — a practice unique among autonomous driving companies. Waymo, Zoox, Aurora, and Nuro all provide detailed narratives in their filings. Tesla does not disclose mileage data, which makes independent crash rate calculations approximate rather than definitive.
Tesla has announced plans to expand its robotaxi service to seven additional U.S. cities — including Dallas, Houston, and Phoenix — in the first half of 2026.
Why Independent Tracking Matters
Autonomous vehicle companies self-report their safety data, and the depth of that reporting varies enormously. Independent trackers fill a critical gap by aggregating publicly available crash reports, NHTSA filings, and news coverage into a single resource. Without them, the public — and potential passengers — would have limited visibility into how these services actually perform.
For anyone injured in a robotaxi incident, these trackers also serve as a starting point for understanding the pattern of incidents and whether the company had prior knowledge of recurring failure modes.
When Police Intervene with Robots
As robots increasingly operate in public spaces, there have been notable incidents where law enforcement has directly intervened — sometimes in ways that raise more questions than they answer.
Macau, March 2026
A Unitree G1 humanoid robot startled a 70-year-old woman on a Macau street at night. The woman was hospitalized for observation after the fright, though no physical contact occurred. Two police officers physically escorted the robot away, footage that went viral as the robot’s “arrest.” Police issued a verbal caution to the operator and returned the robot. No charges were filed.
The incident highlighted the absence of any regulation governing humanoid robots in public pedestrian areas — in Macau or elsewhere.
Lubbock, Texas, September 2024
In a very different context, a law enforcement tactical robot was deployed during a standoff at a Days Inn motel after a 39-year-old man barricaded himself and fired shots at officers. The remote-controlled robot deployed tear gas, removed obstructions the suspect had thrown over it, and physically pinned the suspect to the ground, leading to his surrender. Some media described it as the first instance of a robot physically restraining a suspect in a U.S. law enforcement operation.
The ACLU has raised ongoing concerns about the use of force by police robots, and the Lubbock incident has no established legal framework governing liability if the robot had injured the suspect during the intervention.
What These Incidents Tell Us
These two cases represent opposite ends of the robot accountability spectrum. In Macau, a commercial robot caused unintended psychological harm during routine promotional use. In Lubbock, a tactical robot was deliberately used to apply physical force. In both cases, the legal frameworks for determining responsibility, establishing liability, and protecting the rights of affected individuals are either incomplete or nonexistent.
As robots move from controlled environments into streets, sidewalks, and public interactions, the need for clear legal and regulatory frameworks becomes more pressing. The question is no longer whether robots will interact with the public, but how quickly the law can keep pace.
What Should You Do Next?
If you have been injured or harmed by a robot — whether a robotaxi, a delivery robot, a humanoid robot in a public space, or any other autonomous system — you may have legal options. The law in this area is developing rapidly, and the liable party may include the vehicle operator, the technology company, the deploying business, or the robot manufacturer.
Get a free case review to discuss your situation with a qualified legal professional.
This article is for informational purposes only and does not constitute legal advice. Injured By Robots LLC is not a law firm. Laws vary by state and may have changed since publication. Consult a licensed attorney in your state for advice about your specific situation.