AI for Legal : The story of hype, adoption and what comes next
summary
Generative AI accelerates legal work from TAR-style classification to drafting and research, with retrieval setups like retrieval-augmented generation hooked to firm repositories; large litigations surface millions of documents (one copyright matter cites ~10 million) and single AI summaries can cost $1,000–$2,000 per run. Adoption stalls on six landmines — accuracy, blackbox audits, bias, cloud privacy, empathy gaps, and junior‑lawyer displacement — so practical fixes focus on verbatim citation, immutable audit trails, secure enclaves and a mandatory human-in-the-loop for verification.
AI-written summary · about
chapters (6)
- 0:00Legal AI challenges
- 2:41Hype versus reality in adoption
- 13:43Landmines in legal AI
- 26:25Regulatory compliance in AI
- 34:13Future of legal AI
- 35:15Q&A
related talks
28mEthical AI?!
53mSome Implications for the use of Gen AI in courts
27mAI and Geospatial Analysis. What has changed?
17mGenerative AI: Keeping it real
32mPanel Discussion: Trends in Generative AI
33mAI is coming for your job
35mAalap: An AI Assistant Transforming Legal Functions in India
52mAI Is Just Software, What Could Possibly Go Wrong?