Why People Blame Self-Driving Cars More Than Human Drivers

Participants tended to view the automaker of the not-at-fault vehicle as more responsible when the vehicle was self-driven than human-driven, even though in both situations the crash was unavoidable.

Why? People fixate on the AV itself as an “abnormal” presence, making them more likely to imagine alternative scenarios in which another driver were present instead—but they don’t just imagine any driver, but a “perfect” one. When the researchers asked participants to complete sentences about how things could have turned out differently, they gave answers like “If only there were a person in the car, they would have been able to swerve to avoid being collided into”—even though the accidents were constructed so that such split-second behavior would be impossible.

De Freitas explains: "In effect, these vehicles end up being held to a higher standard than human drivers, because people compare them to an imagined ideal that surpasses what a human is capable of in the same situation.”

Managing risks for AV companies

If the public unduly blames AV companies when they are not at fault, then this could drive up insurance premiums for AV providers, the research suggests. Some companies might have to choose between passing on insurance costs to consumers or reducing their liability coverage, exposing themselves to tremendous risk in the event of technology failure.

“As we saw with the Cruise incident, that would be unwise, because even when it is not at fault, an accident could be potentially devastating for a company’s survival,” De Freitas says.

In fact, the risks De Freitas and team uncovered might be magnified in practice, for at least three reasons:

1. Liability in some US states is uncapped, so juries may award damages beyond actual costs if AVs are unfairly blamed.

2. Even if they are only partially liable, AV companies are seen as wealthier and insured, making them attractive targets of lawsuits.

3. In some states, laws may force AV providers to pay full damages if other parties in the accident lack full insurance coverage.

How to build trust

One upside from the research: The more people trusted AVs, the less likely they were to view them as liable. For these people, AVs may seem less abnormal in the first place, making them less likely to imagine alternative “perfect” scenarios.

A later experiment showed that highlighting the at-fault driver’s responsibility lessened the tendency to fault the AI. By steering attention away from the vehicle’s novelty, people became less prone to exaggerating hypothetical risks.

De Freitas says automakers and other companies expanding in AI need to:

• Anticipate tech failures. “Be aware that these systems can fail, and when they do, you're potentially looking at more risk than usual.”

• Earn consumers’ trust so that the tech seems less abnormal and unsettling. “Explain what you're doing, why it makes sense, why it's reasonable, and what you're continuously doing to improve and be transparent, especially when it comes to safety.” AI-based systems will need to “feel more familiar, like a common part of our surroundings.”

• Rule out counterfactuals. The public will look for what could have happened, even under potentially impossible scenarios. Companies and their legal defendants will need to counter theories with facts, pointing out what AVs can—and cannot—do, and how this limits what counterfactuals are possible.

• Steer attention away from the technology. Focus the public’s attention away from the tech and toward the true sources of fault. Still, redirecting attention is not the same as concealing facts—companies must remain transparent about AI’s role in any accidents.

Image by Ariana Cohen-Halberstam with asset from AdobeStock.