I think legal technology,[1] explains very little of the glut of lawyers. Mostly because it's really not very useful so it doesn't get used that widely.
The real answer is that the job situation for lawyers isn't much different than nearly any other white collar profession. The amount of legal work per unit GDP is pretty steady, GDP isn't growing rapidly, and schools graduate way more students than the industry can absorb.
[1] Obviously lawyers use the same technology that has helped everywhere else. PDF instead of paper copies, email, outsourced IT, etc.
You have a corpus of knowledge that you need to dig through. Cases are interconnected by footnotes (i.e. case X referenced precedent set in case Y [and succeeded]). The relative strength of referencing a case is predicted by the number of references made to it [successfully].
The biggest use of highly-paid, entry-level law grads was to run from case book to case book, chasing down all those footnotes and getting photocopies of them.
Digitizing them (i.e. OCR, make them searchable) and turning footnotes into hyperlinks basically negates the entire point of those entry-level grads.
So, no use of predictive algorithms or fancy ML - just PageRank applied to a set of case books.
Meanwhile, West/Lexis went electronic in the 1970's, and electronic research has been standard at business law firms probably for the last 20-30 years. In any case, the process of legal research was probably faster than you think pre-computers, because cases were all indexed by West/Lexis on the basis of their content. So even back in the days of book research, the primary use of young associates was, as it is today, discovery, not research.
And it's funny you mention PageRank. Google Scholar has gotten better for legal research, but it's still vastly inferior to old-school search queries in WestLaw. PageRank is pretty much useless for legal research. The strength of referencing a case has nothing to do with how often it's cited, but the similarity in the factual/legal scenario in that case compared to your own.[1] A case nobody cites that addresses the specific wrinkle of the law your situation falls into is infinitely more authoritative than one that's highly cited for some generic principle of law.
[1] PageRank is pretty good at finding seminal/controversial cases in an area of law, but that's rarely useful. You already know what those are, because you paid at least a modicum of attention in law school and/or you started your research by reading a chapter in the relevant treatise, which lays those cases out for you.
That said, I'm also imagine part of it is just the same technology that is helping everyone else, albeit in a particularly paper-intensive field that requires enormous amounts of searching and checking.