One thing to note is that a huge amount of legal work done today is actually quite possible to automate. It's a lot of reading through documents and finding the few clauses that actually matter to a contract or document, then focusing in on those. It's defining an overall strategy and then fleshing out (mostly templated) documents to implement that strategy.
Another factor in favor of legal AI is that legal documents have much more restricted grammar than standard English text, leading to less ambiguity. This makes them both easier to parse and to create automatically (you don't need to make the text sound "natural"). By 2030, AI will be able to easily summarize a full legal document and provide relevant references. (In fact, probably much sooner.)
That being said, it's important to note that this report isn't predicting the disappearance of lawyers (nor am I). Instead, it's the collapse of traditional law firms, which depend on a pyramid of partners doing high-level sales and strategy work being supported by associates and paralegals doing the grunt work. The grunt work is rapidly becoming automated, but I don't think AI will ever replace the client interactions and recommendations component. We've seen a lot of success with AI-lawyer pairing and I do think we'll see much more of it in the future. This definitely challenges the organizational structure of law firms. In the future, I think we'll see a lot more small and individual practices where clients work with a partner they trust who can get a lot more done thanks to AI doing all the grunt work.