VAILL

VAILL team members Mark Williams, Emily Pavuluri, and Kyle Turner contributed to this industry-first benchmark project. From the executive summary: "This first-of-its-kind study evaluates how four legal AI tools perform across seven legal tasks, benchmarking their results against those produced by a lawyer control group (the Lawyer Baseline). The seven tasks evaluated in this study were Data Extraction, Document Q&A, Document Summarization, Redlining, Transcript Analysis, Chronology Generation, and EDGAR Research, representing a range of functions commonly performed by legal professionals. The evaluated tools were CoCounsel (from Thomson Reuters), Vincent AI (from vLex), Harvey Assistant (from Harvey), and Oliver (from Vecflow). Lexis+AI (from LexisNexis) was initially evaluated but withdrew from the sections studied in this report." See below for 🔗

Extra shout-out to VAILL team members Emily Pavuluri and Kyle Turner for their contributions to this project!

🚀 Calling all changemakers! 🚀 The justice gap is massive: 260 million legal problems arise each year in the U.S., and over half go unresolved. For low-income individuals, the numbers are even worse—only 8% of legal issues get professional help. We know traditional pro bono alone can’t bridge this gap. But what if technology could help multiply our impact? 💡 Enter HACKING PRO BONO. 💡 On Saturday, April 12, 2025, we’re bringing together lawyers, legal aid leaders, technologists, access to justice advocates, and anyone who’s experienced the legal system firsthand to tackle one big challenge: 👉 How can AI and technology supercharge pro bono legal services? 👈 This is NOT your typical legal event. It’s a high-energy, solution-driven hackathon where: ⚡ Teams will collaborate to build innovative solutions. ⚖️ Legal and tech professionals will join forces to rethink pro bono delivery. 💻 No coding skills? No problem! We need diverse minds to make an impact. 🔥 The best ideas will help real people get the legal help they need. 🎯 How to get involved: ✅ Register your interest now - link in comments! ✅ Form a team or join one at the event. ✅ Help reimagine access to justice! Let’s stop talking about the justice gap and start building real solutions. Who’s in? 🙋♀️🙋♂️ #LegalTech #AccessToJustice #A2J #HackingProBono #AIforGood #ProBono #ProBonoInnovation #JusticeForAll

ICYMI, on the new VAILL Substack *The AI of Law* we've just featured a series of short "AI manifestos" by lab practicum students. Each of the six manifestos offers bite-size insight for law students and legal professionals alike on how to approach genAI. This month we're launching another new thing on the Substack and we think you're going to like it. 🔥 You can check out *The AI of Law* and follow along here: https://lnkd.in/eJM79DU5

What can we learn about AI and the practice of law from *students*? A LOT. Because law students are super-curious, uber-creative, and haven't yet "assimilated." 👊 Check out Ellie Olson's advice in post no. 1 of our series! See comments for link, or check your inbox if you subscribed to The AI of Law.

Starting today, we're sharing a series of manifestos, written by law students for law students. We're calling it "AI goes to law school," and believe us when we say: these students understand more about the impact of AI on the legal profession than many (most?) practicing lawyers. Follow along and learn from the students! https://lnkd.in/dDpW9Gfc

We've seen the LLM-aggregation platform Poe popping up in posts around here and knew we hit on something when we built it into Day 9 of Exploring GenAI. 🔥 Poe is a great way to experiment across GenAI tools and get a good feel for which tools work best for what purposes, and for your specific goals and preferences. Check out Day 9 (it's a great stand-alone lesson) or dive into the entire 10-day journey with us. Link to all the lessons on the VAILL Substack, The AI of Law, in the comments 👇

🏎️ We're cruising right along in our exploration of GenAI! Day 8: Advanced Prompting Techniques 💬 just dropped. Only 2 more days to go, and this first e-course will be a wrap. 🔥 The good news: You can take the course anytime, at your convenience. All 10 days will live on the VAILL Substack, The AI of Law. 🎉 See link to Day 0 of Exploring GenAI with VAILL in the comments. We're adding links to this post for each day's lesson as it drops so you can follow along anytime. Expect to see new learning opportunities around AI for legal professionals in 2025!

ICYMI: The North Carolina State Bar formally adopted its Ethics Opinion on AI in Legal Practice. At the start of this year, 2024 FEO 1 proposed considerations for how lawyers may ethically use artificial intelligence. After two rounds of public comment, this opinion provides high-level guidance on our duties of competence, diligence, and confidentiality. Ultimately, 2024 FEO 1 makes it clear that lawyers can use these powerful tools responsibly. For the full text of 2024 FEO 1, check the Winter 2024 Journal below or here: https://lnkd.in/eAizp6Qx Like most industries, many lawyers are already using AI tools in practice. For those just beginning to explore how AI can support your practice, there are excellent resources available. For instance, Vanderbilt's AI Law Lab (VAILL) just launched a 10-day guided exercise series that will take the guesswork out of getting started with generative AI tools. Check it out here and follow along at your own pace: https://lnkd.in/eDgf3YbY AI is already transforming how lawyers work, and this is just the starting point for ethical guidance. Are there ethical questions about AI in legal practice that you think future opinions should address? #ArtificialIntelligence #genAI #AI #ethics #LegalTech #emergingtech

This post from Ethan Mollick may be his most useful to legal professionals, as we try to figure out the what/how/why of generative AI. He identifies 15 types of tasks that GenAI is useful for, and 5 that are a no-go. The good: 🟢 No. 2 on his list? "Work where you are an expert and can assess quickly whether AI is good or bad." The human in the loop must be able to accurately assess the output. This should assuage fears of lawyers about GenAI taking over our work—or at least most (some?) of it. For now. 🟢 No. 4 on his list: "Work that is mere translation between frames or perspectives." When combined with No. 2, this is an ideal use of GenAI: when we use it to translate complicated legal information for a specific audience. We only know if the translation is a good one if we are an expert. 🟢 Nos. 5 and 8 go to using GenAI as an idea generator or editor/curator when writing, and if you know how to effectively use a GenAI tool for this purpose, it's a veritable goldmine. Figuring this out for your specific purposes takes some time and experimentation. Put in the work, and you will get value out. 🟢 No. 10: "Work where you need a first pass view at what a hostile, friendly, or naive recipient might think." I suggest that many lawyers could use this to be much more effective in written communication, especially email. Based on having received emails from lawyers for almost 30 years now. 🟢 No. 14: "Work where you want a second opinion. Give an AI access to the data and see if reaches the same conclusion." Again, when combined with No. 2, this can be incredibly fruitful. And, again, you need to understand how to ask the AI, and what information to give it, to get value out. And this requires experimentation. Now, for the bad: 🔴 No. 1 on this list is critical: "When you need to learn and synthesize new ideas or information. Asking for a summary is not the same as reading for yourself." And, No. 4: "When the effort is the point." While obvious, the question for us now is how we're going to stop ourselves (and our students, if you're in legal ed) from doing this very thing. Maybe this apparent curse is a blessing. Because one potential future requires us to be much more intentional about how we teach. And if the way we teach now makes it too easy for students to phone it in (and this happens not just with GenAI), then the problem is not just with the students. The problem is also with us. 🔴 No. 2: "When very high accuracy is required." If you're an expert and can vet the output, you'll know if it's accurate or not. That lawyers continue to rely on GenAI and then don't vet the output is vexing but this is a lawyer problem, not a technology problem. 🔴 No. 3: "When you do not understand the failure modes of AI." and No. 5: "When AI is bad." This is why lawyers need to experiment with GenAI. Full stop. Link in comments. 👇