High Technology Attorney, Entrepreneur, and Computer Technologist; Lead counsel in large tech cases; Helped build Web 2 & 3 services that lots of people use.
I spoke yesterday to the Digital Entertainment Group or DEG about NFT legal strategy - here is the video below - I step through numerous weighty legal-tech issues for any NFT project beginning at about the seven minute mark. I hope you find it helpful.
https://t.co/UfH9wyB8H3
Legal AI needs a Lawyer Analysis Metric (LAM): how much human review counts as sufficient supervision of AI for a given task. Review every citation? Wise, but it erases AI’s time savings. Review nothing? Sanctions. How about the LAM for a given contract review redlined by AI? We’re regulating this line without ever measuring it. It will vary by task. As AI improves, LAM should fall — if liability and ethics rules let it.
In AI Legal apps the biggest time gains are in analyzing huge datasets, ESI, or evidence and organizing the results - use cases where the “human in the loop” time to the volume of materials analyzed ratio is very low - a low LAM. The law profession is in need of best practices on the LAM line - not just a hand wave that “legal judgments are reserved for lawyers.” AI and agent workflows are currently making millions of micro legal judgments a day, like what data or evidence from millions of data points to leave out. Output review doesn’t surface that. The legal profession needs an AI safe harbor for this major innovative change or could be seen to be violating many of the over-broad noble-sounding AI usage rules being put into place. Nuance matters.
Legal AI needs a Lawyer Analysis Metric (LAM): how much human review counts as sufficient supervision of AI for a given task. Review every citation? Wise, but it erases AI’s time savings. Review nothing? Sanctions. How about the LAM for a given contract review redlined by AI? We’re regulating this line without ever measuring it. It will vary by task. As AI improves, LAM should fall — if liability and ethics rules let it.
You are 99% correct. 😂 It is good to know the price performance ratio of a model that a lawyer is starting out with. But the RAG, vector stores, hybrid search, playbook design, guardrails, anti hallucination workflows, human in the loop methods etc matter so much that mild differences between the AI models are a rounding error. In terms of speed the rate limiting step is where the line is drawn on the human time and quality carrying out their duty to supervise AI - that is massive compared to the speed of an LLM.
Quinn Emanuel and firms like them are most likely to thrive in the AI era. Litigation involves skill in talking, presenting, and persuading, AI will supplement and enhance that but not replace it for a long time. QE has great legaltech nerds in their firm. AI “native” startup firms will be AI “same” over time as the foundational models are similar and AI legal usage is capped by the “duty to supervise AI” which is the same for all lawyers - if one delegates to AI too much they cross the line. Human in the
Loop is the AI rate limiting step.
We built the dumbest AI Legal app last year that uses your device microphone - I refused to write about it - when we tested it on friends it became their favorite app. Yes, you will feel like you argued in court. You really do. It’s free. Fun LegalTech!
https://t.co/0bZG8efHtp
@mollyisonchain Great to hear! Actually given the complexity of the matters Coinbase faces in the crypto industry AI legal is particularly helpful to your use case. Proper AI training over time can bolster institutional memory on solving thorny legal matters.
@litigationai This is mostly accurate. GCs are incentivized to train on outside counsel inputs. But there also incentives to GCs to have outside counsel be responsible for certain legal work product. Multiple forces at work.
They are likely at the top of the AI focused law firms. I like what I know. But for fun ask: What about their tech makes them unique? If all lawyers have the same duty to supervise AI and not over delegate to it and they use the same ever improving AI foundational models - won’t the “AI first” label just become “AI same” over time?
Here is a free “floor” AI app for lawyers.
We just posted a walkthrough video of LALA on https://t.co/n1zIWXkioR It demonstrates free legal research tethered to Court Listener caselaw, detailed log files, storing the full actual cases, generation of legal work product, agents, sharing of the full AI journey in to a LALA file for sharing with colleagues who can review all the above and pick up the AI work and even chats where the sending attorney left off. It’s no install. We built it to work better with our legal team and contractors and we are sharing it for free.
We just posted a walkthrough video of LALA on https://t.co/7nX4pPdZVV It demonstrates legal research tethered to Court Listener caselaw, detailed log files, storing the full actual cases, generation of legal work product, agents, sharing of the full AI journey in to a LALA file for sharing with colleagues who can review all the above and pick up the AI work and even chats where the sending attorney left off. It’s no install. We built it to work better with our legal team and contractors and we are sharing it for free.
We pondered how do we use the AI legal workflows, chat sessions, fetched caselaw, and work product generated by lawyers, such as contractor attorneys, who work on different platforms? How do we see what they did so we can review it and enhance it by leveraging their prior effort? We built LALA to do that. Portable AI legal workflows in a single file you can share with colleagues. Pickup where they left off. No software install required. Can run locally on free AI. Free for all lawyers who want to use it.
@litigationai Or you can try this for free - a browser based no install AI legal app that allows lawyers to share their full AI agent sessions, legal research, and diligence logs with other lawyers by secure file transfer - LALA.
https://t.co/3Za3XCnMjr