Microsoft President Brad Smith called for stronger artificial intelligence safety regulations in a BBC podcast episode, advocating for global implementation. He argued that AI products are subject to less regulation than washing machines and milk packaging.
Smith put forward several specific recommendations: if an out-of-control AI agent system damages other systems, legal requirements should mandate notification to those affected; AI products should be tested by independent evaluators before release; and autonomous systems should also be equipped with "kill switches" capable of halting their operation.
He described AI as a technology that can serve both as a tool and as a weapon. These measures currently represent Smith's governance proposals, and whether they will evolve into concrete rules depends on future policy developments.