Noubar Afeyan, co-founder and chairman of Moderna, proposed implementing mandatory safety checks for commercial AI agents prior to their market release. He shared these views during the Bloomberg New Economy Forum held in New Delhi on October 13, 2026.
Afeyan distinguishes between large language models, which he characterizes as having capability, and AI agents, which possess intentionality. He argues that agents require specific oversight because they operate autonomously with minimal human supervision.
The investor emphasized that his proposal focuses on the commercial deployment of agents rather than foundational model research. He suggests that the technology sector should adopt safety protocols similar to those used in the pharmaceutical industry.
Drawing a parallel to the thalidomide crisis of the 1950s and 60s, Afeyan warned against waiting for negative outcomes before establishing regulations. He noted that autonomous agents could potentially generate harmful compounds or unreliable data.
