BlueSky Thinking Summary

In an increasingly fast-paced landscape of artificial intelligence (AI), Sergio Rebelo and co-authors from the Kellogg School analyze current approaches and propose a new economic model for the optimization of societal outcomes.

Their research critiques single-minded approaches banning and mandatory testing and finds them to be individually insufficient for managing AI's unpredictable impacts on society.

At the same time, they propose a balanced approach: beta testing with limited liability for developers.

It's supposed to align private incentives with public welfare, accommodate incremental innovations, and leave room for high-novelty AI developments.

Despite these challenges in its global implementation, and pervasiveness of AI across the world, Rebelo provides a basic framework, much like drug approval processes, to ensure responsible AI innovation amidst regulation uncertainty.