He urged, nevertheless, that one of the best precedent right here isn’t FINRA, it’s INPO, the Institute of Nuclear Energy Operations, which the nuclear business created inside months of the 1979 Three Mile Island partial reactor meltdown “on the logic that an accident anyplace is an accident in all places. INPO peer-reviews each US plant, its evaluations transfer insurance coverage premiums, and it sits on high of the NRC’s statutory flooring. That could be a public-private stack very near what Hassabis is describing. Frontier AI has the identical construction: one lab’s catastrophic failure brings regulation down on all of them.”
For enterprise CIOs and different IT executives, Goryunov stated, that mannequin has the potential for being an enormous win.
“Right this moment, each enterprise duplicates the identical AI diligence of red-teaming, eval suites, governance committees and every does so with much less info than any certifying physique would have,” Goryunov stated. “A reputable requirements regime does for AI what UL did for electrical gear and SOC2 did for cloud: it converts an unknowable danger right into a procurable product and offers boards a defensible commonplace of care. That’s not crimson tape. That’s peace of thoughts with an audit path.”
