Recent findings from Alphabet’s Waymo and Tesla suggest that self-driving cars are not yet fully autonomous, with humans still playing an essential role in ensuring vehicle safety. Both companies have shared information about their “remote assistance” services, which intervene when software encounters complex or hazardous situations.
Waymo, known for its robot taxi services operating in six US cities and poised to expand globally, has confirmed that it employs 70 remote assistants monitoring over 3,000 vehicles. Approximately half of these assistants are Filipino contractors who must complete comprehensive training on U.S. traffic rules and hold valid driving licenses.
Waymo Vice President and Global Head of Operations Ryan McNamara highlighted the company’s reliance on a US-based team to handle its most challenging scenarios—such as managing accidents and interactions with law enforcement. The company emphasizes that ensuring safety requires adhering to stringent criteria, including drug and alcohol assessments for all remote assistants.
In contrast, Tesla operates two domestic offices of remote operators—one located in Austin and the other in the San Francisco Bay Area. According to AI technical program manager Duy Cao, Tesla’s human monitors require background checks, active driver licenses, and must work from US locations. Notably, Tesla has reduced its reliance on physical observers inside vehicles but continues to use chase vehicles or operators for safety oversight.
Carnegie Mellon University autonomous vehicle researcher Philip Koopman emphasizes that even the most advanced AI systems cannot foresee every possible scenario encountered on roads. “For the foreseeable future,” he stated, “there will be roles for humans in guiding and ensuring the safety of these vehicles.”
These findings underscore the necessity of continued human intervention alongside self-driving technology, a balance critical to advancing safe autonomous vehicle solutions.