This week we’ll be sharing more takeaways and hot takes from the Legalweek conference in New York last week. First up is is the very first AI workshop of the conference. Billed as a group of visionaries (we’d agree with that) discussing the state of AI across the legal market, perhaps the most interesting part was when they answered some burning audience questions.
Danielle Benecke, the founder and global head of Baker McKenzie’s machine learning practice, said that if 2023 was the year of discovery, 2024 was experimentation and 2025 was where we started to see some deployment in legal workflows. This year is where value – the economics and accountability of AI – will be under the microscope. Incidentally, next year, Benecke says we will be “blown away.”
Baker McKenzie has had a significant focus on Copilot and ‘systems that can produce outcomes’. The latest models appear in the firm’s software and cloud platforms, and the constraint is not whether they have access but how they steer it: interesting observations as the industry wrestles back and forth with whether the dominant approach going forward will be a Swiss pocket knife.
Legal is being pulled in multiple different directions but the next phase will be defined by who is building the most trusted workflows, said Benecke.
Clients and partnership were a big theme of the conference. Zeynep Ersin, chief innovation & strategic design officer at Seyfarth Shaw said that there is an expectation now of a faster turnaround without compromising quality, and the expectation that you will partner on clients on their own GenAI journey – that is coming through strongly in RFPs and how firms are being evaluated.
There is a risk around not using AI now. Ilona Logvinova, chief AI officer at Herbert Smith Freehills Kramer, said that if you’re not using AI on a data heavy exercise, you’re not using the best tools. “It’s critical that we offer clients optionality and that law firms are fluent in multiple tools,” she said. This hasn’t necessarily been true for law firms but major consultancies such as McKinsey meet the client where they work. “It’s up to us to think how we create that opportunity for them and to think about the value proposition,” she observed.
Efficiency is still a nuanced conversation – oftentimes efficiency translates to less hours and a lower price, but, Ersin pointed out, it may also mean achieving higher outcomes and identifying risk that we couldn’t before. What is key is to be transparent with clients and build a relationship of trust.
Benecke’s team charges on an entirely fixed fee (with some flexibility) basis. The value is around the scoping and framing and understanding the risk. Clients want someone else to assume liability if something goes wrong, and that is part of the value transfer, she said.
Logvinova observed that it’s a bit of a red herring to talk about whether the billable hour stays or goes. The likelihood is that it will stay but will it be a measure of time or value, she questioned, observing again: “That is for law firms to analyse.”
For Logvinova, too much responsibility is being put in the laps of clients, instead of law firms rolling up their sleeves and solving the problems themselves. Keep an eye out for our interview with her on that and other topics.
Some of the most interesting and arguably explosive points from the panel, ably chaired as always by LexisNexis’ technology chairman Jeff Reihl, came in response to audience questions over Slido.
How do you train people when AI tools change so quickly? Logvinova said that that tools need to be intuitive. “Once you understand how to navigate them, it should be transferrable.”
If your firm is radically more efficient, how is that being passed on? Benecke said that the firm has big, global, five-year retainers that reward efficiency and where some of the work is based on playbooks. Increasing efficiency and speed frees up budget to get to higher value work. The firms also has an “AI native” cyber incident response offering that would be impossible without technology.
How important are LexisNexis/Westlaw and can they be replicated? This is in response to the legal software market crash following the launch of Anthropic’s Claude legal plug in. No doubt to Riehl’s joy, Oliver Roberts, a professor and managing attorney at The Roberts Legal Firm, said: “No-one can replicate that data.” Harvey isn’t grounded, he said. While it has a Lexis overlay, lawyers need access to the case law, the headnotes, the classification. “That’s hard to replicate and absolutely vital,” Roberts said.
Are Harvey and Legora just hype? Logvinova said: “Definitely not.” These platforms are tailored to legal and assistive in a way that is a step change to how the broader based large language models are working. “It’s just a question of how they fit in your organisation’s strategy,” she said.
In 2030, what will we look back on and say we underestimated? This was one of the most talked about moments of the panel or even the conference. Roberts said: “It’s the biggest misstatement that AI won’t replace lawyers but will replace lawyers who use AI. It will replace them.” While this may be a hot take, it was also the first time that many people in the audience had heard someone say it in open forum.
The post Legalweek: The state of AI, red herrings, and end of lawyers appeared first on Legal IT Insider.