There was a strong perception that AI developers should be held responsible for crashes caused by self-driving vehicles.

Electrek reported on Sept. 27 that about 90 percent of respondents in a survey on who should be responsible for self-driving crashes said the developers of the AI that moves the vehicle are responsible. About 3,000 people took part in the survey.

The survey was conducted against the backdrop of a series of self-driving issues that recently drew attention. The question cited Tesla’s repeated fatal crashes, Waymo’s large-scale parking violations and the expansion of robotaxi operations, and asked: “Who should be responsible when a taxi driven by AI fails?”

The option that drew the fewest votes was personal liability for the person sitting in the driver’s seat. The choice that said, “You are the one who called the taxi, and it is common sense that human intervention is sometimes required, so that human is you,” received 102 votes. That was about 3.5 percent of the total. It is an argument that the passenger bears ultimate responsibility for intervention even if autonomous driving is operating, but it drew limited support in the responses.

Many respondents also said vehicle owners should be held responsible. Some argued it is excessive to impose legal responsibility on manufacturers or technology providers, noting that any car can be used for speeding, illegal parking or even harming others. One respondent said, “All cars have the ability to break the law,” and argued it is not reasonable to pin such cases on the manufacturer.

There were also views stressing AI developers’ responsibility. Steven Fields (스티븐 필즈) argued it should be examined what measures a self-driving car owner can actually take to prevent an accident. If the owner fails to follow manufacturer-recommended software updates or comply with operating areas, the owner can be held responsible. But if there was no way to prevent the accident, responsibility should not be assigned, he suggested.

Some said responsibility should not be attributed to a single party. They argued it should be shared because vehicle manufacturers and AI developers build autonomous driving systems together. The view was raised that judging crash responsibility should also consider the degree of control held by the manufacturer and AI developer, and whether the accident could have been prevented, not just choices made by owners and users.

The timing of the survey also coincided with a recent self-driving debate. Tesla re-unveiled the Cybercab earlier this month in Austin, Texas, and the Tesla Semi, whose production had been delayed, began mass production in Nevada. The survey was conducted as issues involving robotaxis and commercial electric trucks drew attention at the same time.

The survey shows that as self-driving services expand, disputes over crash responsibility could also grow. In particular, as robotaxis operate in real city centers and companies expand services built on autonomous driving technology, establishing standards for determining who is responsible when accidents occur is becoming more important.

The survey showed a split between views that technology companies should bear responsibility in line with the level of safety control they hold, and views that responsibility should also consider choices made by vehicle owners and users. As autonomous driving spreads, how to allocate responsibility when accidents occur, as well as the safety of the technology, is emerging as a key issue.