Seat Belts For AI

A year ago, along with 30,000+ other people, I signed the Future of Life Institute (FLI) petition to pause giant AI experiments … exactly how much effect it had is obviously impossible to say - the experiments are still ongoing, so maybe it made zero difference.

But maybe - maybe - that petition made a difference by helping to raise awareness of the risks associated with AI.

Deepfakes are just one aspect of this. Artificially created voices, images or videos, that any reasonable person could easily mistake as being real.

OpenAI just announced Sora, their text-to-video tool: type a few words, and it creates astonishing videos. If you’re not aware just how good they’ve got, have a look here - or here.

To be clear: I’m not ‘anti-AI’. I actually think that’s an impossible position to hold - AI is here, and it’s not going away. The question now is whether or not regulation can keep up. We have seat belts and speed limits on roads for good reasons. We need equivalent laws around AI, because business does not have a good track record for self-regulation.

[I’m looking at you, tobacco, fossil fuel and social media companies].