on the contrary, it may help to highlight incongruities in the legal domain and provide lawyers with compelling groundwork to make relevant claims

Or you could take the view that, in fact, this is one of the things LLMs are very good at, ie making sense of complexity.

But the lawyers reading the law won't be the only ones using LLMs. LLMs will also be used to write laws. Then lawmakers will use them to "check" that their 20,000 page law supposedly works. No human can understand the scope of today's laws: how much less when no LLMs can understand the laws created 20 years from now.

I'd love to hear that LLMs can be used to trim and simplify complexity, but I don't believe it. They generate content, not reduce it.