The best way to counter bad artificial intelligence is using good AI

​iI’s critical that policymakers avoid focusing too narrowly on catastrophic risk and inadvertently hamstring the creation of positive AI tools,.