Tech Giants Agree to Voluntary AI Safety Checks by Governments Before Public Release

Top tech firms like Meta, Google, and OpenAI will allow governments to vet their AI tools before public release, announced by UK PM Sunak at AI safety summit.

Sunak says AI poses grave threat to humanity, likening it to nuclear war; convened diverse summit including Elon Musk.

Firms agree to test AI models with governments against dangers like national security and societal harms.

Sunak announced international support for expert AI safety panel inspired by IPCC, to be chaired by Yoshua Bengio.

Agreements reached reduce AI threat says Sunak, but testing is voluntary as legislation would be too slow.