The problem with the EU regulation is the same as always, first and foremost they do not understand the topic and can not articulate a clear statement of law.

The problem with the EU regulation is the same as always, first and foremost they do not understand the topic and can not articulate a clear statement of law. They create mountains of regulations, which are totally unclear and which require armies of lawyers to interpret. Adherence to these regulations becomes a major risk factor for all involved companies, which then try to avoid interacting with that regulation at all. Getting involved with the GDPR is a total nightmare, even if you want to respect your users privacy. Regulating AI like this is especially idiotic, since currently every year shows a major shift in how AI is utilized. It is totally out in the open how hard training an AI "from scratch" will be in 5 years. The EU is incapable of actually writing laws which make it clear what isn't allowed, instead they are creating vague corridors how companies should arrive at certain outcomes. The bureaucrats see themselves as the innovators here. They aren't trying to make laws which prevent abuses, they are creating corridors for processes for companies to follow. In the case of AI these corridors will seem ridiculous in five years.