Meta's $18B Settlement Signals Need for AI Regulation

Meta agreed in September to pay $18 billion to settle claims tied to AI-driven harms on its platforms, one of the largest sums any tech company has paid out over algorithmic products. The figure landed in the middle of an argument that's been running through Silicon Valley all year: whether AI needs its own set of laws, or whether existing rules and market pressure are enough to keep companies in check. Jensen Huang, who runs Nvidia, the company that makes the chips powering most of the industry's large models, took the second position at Salesforce's Dreamforce conference on Tuesday.

His company has the most to lose from regulation that slows AI deployment, and the least direct exposure to the kind of consumer harms that produced Meta's settlement. That gap matters. Lawmakers in several states have pushed bills this year targeting chatbot safety, deepfakes, and algorithmic discrimination, arguing that self-policing hasn't worked.

Huang's comments at Dreamforce put him squarely against that push, and against researchers inside AI labs themselves who've warned that the technology behaves in ways its own builders don't fully understand.