The EU is a lot more prescriptive than the UK on things like giving people notice of using AI, opt-out for automated decision making, and how the data is used.
UK companies already have to comply with those if they deal with the data of any EU citizen, but if they're UK only it's a bit lighter touch (although still regulated). That makes working with those countries more expensive/more limiting/less attractive, and if we take on those regulations for UK citizens then we will in turn be less attractive to others eg in the US. In the most extreme case, they might restrict our use of their models.
There's a huge race world wide to use AI to improve productivity, with the risk of whole industries being priced out and diminished. (It's more to do with the use of the AI models than with creating them - but obviously, access to models will be critical for that, hence the push for us to develop our own.)
As for the consequences, the AI revolution is a bit like the Industrial Revolution or the uptake of Internet connectivity between companies. We benefitted hugely from being innovators and early adopters in those - it's the kind of difference that sets a country's position in the world for generations.
So putting in place excessive regulations that stop private companies from investing and developing that AI capability is a bit like overly-regulating steel production in Victorian Britain. The innovation will take place elsewhere, and that's where the generations-long advantage will stay.
The other thing that could really scupper the UK with AI is access to electricity. But that's a separate problem.
The UK is actually very well positioned to lead on this. It plays exactly to our strengths: technical innovation, strong education, a flexible and rich private sector who are keen to invest where it will genuinely make a difference, a top flight financial and professional services sector. It's ours to lose... and tying ourselves to excessive regulations from the less innovative EU is one way to do that.