The latest issue of the transportation column, citing data reported by the US Occupational Safety and Health Administration (OSHA), states that autonomous driving test drivers in Waymo and Zoox suffered more than 24 injuries between 2024 and 2025. Most of these injuries were related to sudden braking or abrupt movements of the vehicles, and some employees were unable to work for several months due to issues such as neck sprains.

This report focuses on the less discussed aspect of the expansion of robot taxis. In the past, the outside world has paid more attention to the impact of autonomous driving systems on passenger and road safety, while the occupational risks faced by test personnel and those accompanying the robots are not commonly discussed in public forums.

Injuries are often related to sudden braking.

The report mentioned that the data submitted by Waymo and Zoox to OSHA recorded several work-related injury incidents caused by autonomous vehicles suddenly braking, making emergency stops, or performing abnormal actions. In some cases, the test drivers experienced significant impacts inside the vehicles and subsequently required a considerable amount of time to recover.

TechCrunch also interviewed current and former employees to supplement information beyond what is publicly reported. The report suggests that these records indicate that even if no public traffic accidents occur with autonomous driving systems, they may still pose ongoing risks to the personnel inside the vehicles during testing and operational phases.

More companies or those that have not disclosed fully...

The report also points out that this issue may not be unique to Waymo and Zoox. Test personnel from other autonomous driving developers may also encounter similar situations, but some companies may not fall within the scope of mandatory reporting required by OSHA, making it difficult for the outside world to obtain complete data.

This means that the currently publicly available figures for injuries are more likely to represent the disclosed samples rather than the overall picture of the entire industry. As robot taxis and unmanned delivery vehicles undergo expanded testing and commercial operations, related labor safety issues may receive more regulatory attention.

Waymo mentions sensors and computing power

In addition to security issues, the article also mentioned that Waymo recently summarized its experience after its autonomous vehicles have accumulated over 200 million miles of autonomous driving. Waymo believes that multi-modal sensors are still indispensable, a judgment that contrasts with Tesla's preference for a pure vision-based approach.

Waymo also mentioned that the role of visual language models in autonomous driving is on the rise. At the same time, the company's previously disclosed 5-nanometer ASIC has also attracted external attention. Waymo later clarified that the machine learning computing power of over 1,000 TOPS refers to the entire front-end processing system, not the performance of a single chip.