I really struggle to think of what sort of legal tech would help you interview a poor client that may not speak english better than a good grasp of Spanish and a pad of paper. These are not complex cases with tons of moving parts. As for legal research--your example of finding factually on-point cases is exactly what existing search algorithms do the worst job of.
Maybe I'm pessimistic. But the gap between promises and actuality in the legal tech field is very wide, at present, so I'm really skeptical when I hear about how it will help solve what is really a political/social problem.
Well, if the public defender doesn't know Spanish using software which helps them to understand their client is a start.
I definitely agree with you on just how bad existing legal research software is at finding factually on-point results. This was part of my motivation to help found ROSS Intelligence, which is building better legal research software with ML and NLP at its core.
And while pessimism and skepticism does help identify the difficulties in implementing different solutions, it's optimism that helps find the opportunities imbedded in the difficulties found :)
Maybe I'm pessimistic. But the gap between promises and actuality in the legal tech field is very wide, at present, so I'm really skeptical when I hear about how it will help solve what is really a political/social problem.