Artificial intelligence has the potential to revolutionize healthcare. However, many physicians are still reluctant to put their faiths in AI, fearing liability over medical malpractice. 

Now, one study suggests that physicians who receive advice from AI systems can actually reduce the risk of liability, not the other way around.

In a study entitled When Does Physician Use of AI Increase Liability published in The Journal of Nuclear Medicine, researchers Kevin Tobia, Aileen Nielsen, and Alexander Stremitzer investigated the degree of liability doctors have when taking to AI for medical treatment recommendations.

To determine potential jurors’ judgments of liability, the researchers went online and devised an experimental study. With a representative sample of 2,000 U.S. adults, the researchers gave participants a total of four harm scenarios to which the participants assessed the physician’s liability.

The researchers hypothesized that lay judgments of liability are driven by two clauses: whether the treatment was AI-recommended, and whether the treatment followed standard protocols.

Better off with AI

Results from the online experiment followed suit with the researchers’ hypothesis.

“If physicians receive a standard-care AI recommendation, there is a legal incentive to accept it,” the researchers revealed.

“All else being equal, participants judge accepting a standard-care recommendation as more reasonable than rejecting it,” they added.

On the other hand, the rejection of nonstandard AI recommendations does not diminish physician liability. That being said, physicians are generally put in a less favourable position when giving nonstandard advice.

Overall, the researchers found that accepting AI advice can actually shield physicians from liability.

“The tort law system is unlikely to undermine the use of AI precision medicine tools and may even encourage the use of these tools.”

“Contrary to the prediction of those legal theories, the experiments suggest that the view of the jury pool is surprisingly favorable to the use of AI in precision medicine,” the researchers concluded.

Image credits: Andrea Piacquadio | Pexels