White House wants to vet powerful AI models for risks − a computer scientist explains why AI safety is so difficult

Once powerful AI models are released, it’s nearly impossible to keep them from being misused. Minimizing the risk means ...