Robotaxis, or self-driving vehicles, have made significant strides in recent years, but they still face a range of challenges. While companies like Google's Waymo claim their autonomous vehicles (AVs) are significantly safer than human drivers—reporting a 65% reduction in car crashes and a 94% drop in serious injuries—some experts argue that more testing is needed to confirm these safety improvements. A recent report by the Insurance Institute for Highway Safety (IIHS) supports these claims, but skepticism remains due to conflicting data and ongoing safety concerns. A September 2026 report by TechCrunch highlights some of the risks involved in training autonomous vehicles. According to data from the Occupational Safety and Health Administration (OSHA), companies like Waymo and Zoox—owned by Amazon—caused injuries to test drivers during the training process. These injuries, often involving sprains and whiplash, were linked to sudden stops and other errors as the AVs learned to navigate urban environments. However, the dataset is incomplete, as OSHA's reporting requirements are limited. Inside sources suggest that similar issues continued at Zoox into 2026, though details about Waymo’s latest tests remain unclear. Despite the promise of safer driving, some AVs have shown troubling behavior during their real-world deployment. For instance, Tesla’s robotaxis recorded more accidents than human drivers during their rollout in Austin, Texas. Reports have described incidents such as vehicles plowing over traffic barriers, swerving into oncoming traffic, and running red lights. While these issues are rare, experts warn that they could become more frequent as AVs become more common on the roads. Autonomous vehicles also struggle with unpredictable or unusual road conditions. In July 2026, Amazon’s Zoox recalled over 100 vehicles after they failed to detect smoke, leading to an incident where a vehicle drove into an active emergency. Similarly, Waymo issued a recall after several of its taxis drove into flooded areas. Tesla, which recently launched its Cybercab robotaxi service, is under federal investigation for issues with its self-driving software in low-visibility conditions. Other concerns include AVs driving into closed roads, violating school bus safety rules, and interfering with emergency responders. Despite these challenges, passengers can take steps to stay safe when using robotaxis. If a vehicle swerves unpredictably, speeds excessively, or makes sudden stops, it may be a sign of a problem. More subtle issues, like taking overly complex routes or missing stops, could also indicate technical problems. If a user notices glitches in the app or the vehicle’s display, that could be a sign of a software issue. Most robotaxi services allow users to end their rides if they detect a problem. For example, Waymo offers a “Pull Over” command in its app and in-car display, which directs the vehicle to a safe stopping point. Tesla’s Cybercab service provides a similar feature through its app and touchscreen, and Zoox has an “Emergency Call Button” in the overhead area. Knowing how to use these features can help ensure a safe ride. If an incident occurs, passengers should treat it like any other car accident, following standard safety procedures and contacting emergency services if necessary. Robotaxis typically contact their support teams automatically, but users should be prepared to reach out directly if needed.