Well, maybe using strong language is actually appropriate now, given that no one can really estimate the risk and most experts agree that even total extinction is a possibility.

For me that's enough reason to simply stop this technology. A higher than 0.1÷ risk of killing 1 million children would also be enough.

And my suggestion for regulation would be: hold AI developers 100% accountable for anything their system does.

Calculate the fines based on market cap and harm created. And hold executives personally to account for anything that was irresponsible to release. Including life long prison, if people die from these systems actions.

Let them decide how much risk they want to take. But let it be risk for themselves, not for everyone else who can't afford a private island.