Government-Backed Insurance for Artificial Intelligence Technologies

Artificial intelligence (AI) is an unpredictable technology that has the capacity to both help and harm people. Although insurance plays a key role in compensating for harms in other contexts, AI-produced damages evade traditional principles of risk pricing which limits viable commercial insurance coverage. AI requires modified insurance systems that can compensate diverse and unpredictable losses. Just like AI, at one time nuclear energy was viewed as a new and profitable, yet wholly unpredictable, technology that had the capacity to cause devastating harm.

AI poses similar threats to society in certain domains, including, for example, health care (e.g., risk management tools) and transportation (e.g., autonomous vehicles). This Article explores the challenges of quantifying AI harm and providing insurance coverage for such harms. This Article proposes insuring emerging AI-enabled technologies through a government- backed insurance paradigm similar to the Price-Anderson Act, which Congress created to respond to threats related to nuclear energy. It develops a framework and a pricing model, and it proposes the necessary oversight required to allow AI to continue to progress while simultaneously compensating victims that are vulnerable to the harms caused by the evolving technology.