In brief
- Waymo told senators that human operators overseas provide guidance to its vehicles in difficult situations but do not control driving.
- Sen. Ed Markey said overseas involvement raises safety, cybersecurity, and accountability concerns.
- Waymo defended the practice as limited contextual assistance as Congress weighs new federal AV rules.
Waymo’s claim that its vehicles operate without human drivers drew sharp scrutiny on Wednesday after a senior executive acknowledged that overseas operators help guide the company’s autonomous cars through difficult situations on U.S. streets, raising new questions about what “self-driving” really means.
Testifying before the U.S. Senate Committee on Commerce, Science, and Transportation, Waymo Chief Safety Officer Mauricio Peña confirmed that the company relies on remote human operators to assist vehicles when they encounter scenarios they cannot independently resolve.
“They provide guidance, they do not remotely drive the vehicles,” Peña told lawmakers. “Waymo asks for guidance in certain situations and gets an input, but the Waymo vehicle is always in charge of the dynamic driving task.”
The answer did not satisfy Senator Edward Markey (D., Mass.), who asked whether all of the remote operators were based in the U.S. Peña said they were not, later identifying the overseas location as the Philippines.
Markey warned that overseas human involvement undermines public claims of full autonomy while introducing safety and cybersecurity risks.
“Having people overseas influencing American vehicles is a safety issue,” Markey said. “The information the operators receive could be outdated. It could introduce tremendous cybersecurity vulnerabilities.”
Lawmakers also questioned whether remote operators abroad are subject to U.S. licensing or regulatory standards, and criticized Waymo for outsourcing what he described as one of the few remaining human jobs in an increasingly automated system.
National framework
The Senate hearing comes as Congress weighs whether to include a national autonomous-vehicle framework in the next surface transportation reauthorization bill, and as Waymo faces additional scrutiny following its attempts to scale its business into new markets.
On Monday, the robo taxi company announced a $16 billion funding round, bringing the valuation of the company to $126 billion.
In an interview with Decrypt, a Waymo spokesperson clarified that the company does not consider its remote operators to be drivers and rejected the idea that humans are controlling vehicles in real time.
“Their role is not to drive the vehicle remotely. They’re not remote drivers,” the spokesperson said. “They answer, generally speaking, multiple-choice questions posed to them by the vehicle.”
“All of the driving actually happens on board of that vehicle. It doesn’t happen remotely,” they added.
The spokesperson said both U.S.-based and overseas fleet response agents are licensed drivers and receive training on local road rules in the regions they support.
“All fleet response agents, both those based in the U.S. and those based abroad, have vehicle or van driver’s licenses,” the spokesperson said. “They learn about local road rules where they will be providing remote assistance.”
Waymo characterized the human input as contextual rather than directive.
“The human offers a suggestion in a challenging scenario, and the Waymo Driver will take that suggestion into account when making its next decision,” the spokesperson said.
Some safety researchers said remote human involvement can still play a decisive role when things go wrong.
Philip Koopman, a professor emeritus at Carnegie Mellon University who studies autonomous-vehicle safety, pointed to past crashes in which remote assistance contributed to errors.
“Even though Waymo says these remote assistants aren’t actually steering or braking, their helping out can substantially contribute to a crash, and, in my mind, that makes them a backup driver,” Koopman told Decrypt.
William Riggs, a professor at the University of San Francisco who studies autonomous-vehicle policy and deployment, said the use of off-site human support is consistent with how autonomous systems are designed and regulated today.
“It is also important to distinguish between ‘remote human assistance’ and ‘remote driving,’ as these terms are often conflated,” Riggs told Decrypt. “Remote driving involves directly controlling a vehicle from a distance, whereas remote supervision—used by companies developing true Level 4 autonomy—enables vehicles to make on-board decisions independently, with remote supervisors acting as a support layer rather than direct operators.”
“The vehicles still operate within their agreed-upon and permitted Operational Design Domains (ODDs),” he added. “These ODDs are carefully defined in collaboration with state regulators and adhere to evolving legal frameworks around driving rules.”
The hearing comes as U.S. states continue to issue regulations governing autonomous vehicles and as Congress weighs whether to pass a federal law overseeing the technology. Waymo and other companies, including Tesla, which also testified during the hearing, face heightened scrutiny over their self-driving capabilities.
Daily Debrief Newsletter
Start every day with the top news stories right now, plus original features, a podcast, videos and more.
Source: https://decrypt.co/357128/waymos-self-driving-claims-reality-check-capitol-hill
