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 making safety the top priority in developing the systems.