Successful AI in radiology and beyond requires a new standard for legal liability

Radiology is arguably at the forefront of artificial intelligence in medicine. But for AI to succeed in imaging and beyond, experts say a new approach to legal liability is needed.

One of the simpler ways to do this is by adopting and changing standards of care, Penn Medicine and Harvard experts argued July 13 in the Harvard Business Review.

For example, radiologists could lean on AI to provide initial image interpretations and then perform a subsequent secondary read, the group explained. Many studies have reported successful outcomes using this approach, helping rads detect coronary artery disease and avoid overlooked wrist fractures on X-rays.

“Once this becomes a standard of care in radiology practice, then the potential liability burden of AI becomes less on the individual physician if he or she complies with the particular standard of care,” George Maliha, MD, a second-year internal medicine resident at the University of Pennsylvania Health System, and co-authors wrote July 13.

Certain cases, however, such as injury liability, will be more difficult to handle, the group noted. One solution may be specialized tribunals, which simplify issues and can exist alongside traditional liability systems. It’s feasible for such a system to adjudicate specific algorithms or accidents resulting from them, the group wrote.

Ethical and legal issues related to AI in radiology have been well covered. While most imaging providers believe such tools will enhance their profession, more than 60% say these problems will hinder AI implementation, according to findings from a large international survey.

Read the entire piece in Harvard Business Review below.