Regulating "algorithms" doesn't mean banning. If an algorithm is used to make a decision, a human should be in the loop. There is no way these smoke detectors have any real algorithmic complexity, it's just a vague word used to hide intent and responsibility. How the EU AI agreement, focused on copyright and security, relevant to this?

I can write an "algorithm" that uses hotel data to determine if you should be charged a penalty. Now we can't question it because it was an algorithm.

> def charge_extra(data): return True