This article seems like complete bullshit.<p>I don't think we're going to do anything of the sort and I don't think this whole idea that we are somehow over-regulating anything has anything to do with reality. The people who want to remove this 'over-regulation' are American firms, primarily focused on tracking and advertisements.<p>I don't find anything concrete in the article that we are to have rolled back. It sort of tries to conflate the EU and Europe to get some kind of 'win', but Britain is its own thing, and it then tries to claim that we're rolling back regulations.<p>This will <i>never</i> happen because it's idiocy and we all know it. It's pretty obvious when the only ones arguing for it are foreign competitors. There are AI rules that can probably be loosened-- on model sizes etc., but it's not like people can serve up arbitrarily large models to people anyway.
Europe needs to roll future refinements to AI regulation into the existing GDPR statutes and not label them as AI. What really matters is HOW info is misused or what physical activities were injurious, not whether the miscreant is automated. Rather than prevent harm through proscription of possible causes, both physical and synthetic, future regs need to prohibit the outcome itself regardless of the cause. As I understand it, that's what GDPR does.