I guess the issue is the definition of problem. If the law reaches a point where you need AI augmentation to understand it, that's a good sign it's being applied to problems it can't actually address or which may not even exist at all.

Look at GDPR. It's impossible to know what it really means. Huge efforts are put into action with no idea of whether it will be considered good enough or not. That's not a problem you can fix with ai. You need better law (in this case, no law would be better)