Key takeaways:

• • • Violence against judges is escalating 鈥� Targeted shootings, coordinated harassment campaigns, and threats that now routinely follow judges to their homes and families.

    • The rhetoric driving the escalation is coming from the highest levels of government 鈥� The absence of any public denunciation from the Department of Justice is highlighting the source of the problem.

    • Will the violence itself become part of judicial rulings? 鈥� The endgame of judicial intimidation isn’t that judges stop ruling, it’s that the threat of violence becomes a silent presence in the deliberation itself.

NEW YORK 鈥� Those attendees who came to the recent听 to talk about AI, agentic workflows, and the business of legal technology, also were treated to a session that will likely stay with attendees and had nothing to do with AI.

In that session, four federal judges took the stage; but they were not there to talk about pricing models or AI adoption. They were there to talk about staying alive.

Setting the stage

Jason Wareham, CEO of IPSA Intelligent Systems and a former U.S. Marine Corps judge advocate, introduced the session 鈥� a panel of four sitting United States District Court judges 鈥� by speaking of how the rule of law once seemed resolute, yet how that faith in that has been shaken, year after year. He worked hard to frame his observations as nonpartisan, a matter of institutional fragility rather than political allegiance. It was a generous framing, but it was one that would not survive the weight of the ensuing discussion.

The Honorable Esther Salas of the District of New Jersey said that the reason she was there has a name. On July 19, 2020, a disgruntled, extremist attorney who had a case before her court arrived at her home during a birthday celebration. He shot and killed her twenty-year-old son, Daniel Anderl. He shot and critically wounded her husband. She has spent the years since on a mission to protect her judicial colleagues from the same fate.

The new normal

Next, the Honorable Kenly Kiya Kato of the Central District of California described what has changed. Judges鈥� rulings are still based on the Constitution, on precedent, and on the facts; but what’s different is the small voice in the back of a judge’s head. That voice, often coming after a judge issued a decision that they now have to fight against, asks: What will happen after this? It is now expected, Judge Kato explained, that a high-profile order will bring threats. When two colleagues in her district issued prominent decisions, her first thought was for their safety. That is not how it has been historically.

The Honorable Mia Roberts Perez of the Eastern District of Pennsylvania asked how we got here, pointing to language from the highest levels of government: judges called monsters, a U.S. Department of Justice declaring war on rogue judges, and recently politicians bringing justice鈥檚 families into the conversation.

Judge Salas pushed even further. She acknowledged the instinct to frame the problem as bipartisan, but said the current moment is not apples to apples. It is apples to watermelons. The spike in threats since 2015, she argued, traces directly to rhetoric from political leaders using language never before deployed against the bench.

The federal judiciary is looking to break annual records for threats [against judges], and there is an absence of any public denunciation from the Attorney General or the DOJ.

The evidence is not abstract, nor are the victims, and the panel walked through it. Judge John Roemer of Wisconsin, zip-tied to a chair and assassinated in his home. Associate Judge Andrew Wilkinson of Maryland shot dead in his driveway while his family was inside. Judge Steven Meyer of Indiana and his wife Kimberly, shot through their own front door after attackers first posed as a food delivery, then returned days later claiming to have found the couple’s dog. Judge Meyer has just undergone his fifth surgery since the attack.

All of these incidents happened at the judges’ homes.

Judge Salas then played a voicemail, one of thousands that federal judges receive. It was less than 30 seconds long, but it did not need to be longer. While names had been redacted, what remained was a torrent of threats and obscenities, graphic, sexual and violent, delivered with the confidence of someone who does not expect consequences. Some judges receive hundreds of these after a single ruling, often from people with no case before them at all.

The shadow over the courts

Throughout the session, there was a presence the panelists circled but rarely named directly. A shadow that shaped every observation about escalating threats, every reference to rhetoric from the top down, every mention of language never before used by political leaders, of action or inaction the likes of which would have been unthinkable just several years ago. The specifics were spoken. The name, largely, was not.

It didn’t have to be.

Judge Kato said that what was perhaps the most disheartening aspect of all this is that these threats are getting worse. The people who know better are not doing better. Indeed, she said her children think about these problems every day. What will happen to mom today? Will someone come to the house? These are questions children should not have to carry. They did not sign up for this, and neither did the judges.

In 2026, Judge Salas noted, the federal judiciary is looking to break annual records for threats. She also noted the absence of any public denunciation from the Attorney General or the DOJ. The silence, she said, says a lot.

Not surprisingly, the implications extend beyond the judges themselves. As Judge Salas noted, if judges have to weigh their safety alongside the law, ordinary people don’t stand a chance. If one party is stronger, better funded, or more willing to threaten, then the scales tip.

That is the endgame of judicial intimidation. It鈥檚 not that judges stop ruling, but that the violent and the powerful 鈥� indeed, the people least fit to hold the scales 鈥� can tilt them at will.

That concern echoed an earlier warning from Judge Karoline Mehalchick of the Middle District of Pennsylvania. Judge Mehalchick said that judicial intimidation feeds on misunderstanding. When the public no longer grasps why judges must be insulated from pressure or conversely, mistakes independence for partisanship, the threat environment becomes easier to justify, easier to ignore, and harder to reverse.

What is perhaps the most disheartening aspect of all this is that these threats are getting worse, and the people who know better are not doing better.

In his 2024 year-end report, U.S. Supreme Court Chief Justice John Roberts identified four threats to judicial independence: violence, intimidation, disinformation, and threats to defy lawfully entered judgements. The panel discussed this report as prophecy fulfilled. Public confidence in the judiciary has plummeted since 2021, and the reasons are complex. The judges insisted they are still doing their jobs the right way, but the violence is spreading anyway.

What survives

Judge Salas asked the audience to watch their thoughts. Are they negative and destructive, or positive and uplifting? Can we start loving more? She ended by sending love and light to everyone in the room.

The judges were visibly emotional on the stage.

The words were beautiful. They were also, in the context of everything that had just been described 鈥� the killings, the voicemails, the zip ties, the pizza deliveries masking a threat under a murdered son’s name 鈥� resting in a shadow that no amount of love and light could fully dispel on their own.

The room responded with a standing ovation.

Thousands of people came to Legalweek 2026 to talk about the future of legal technology. For one morning, four judges reminded them that none of it matters if the people charged with administering justice cannot do so safely.

So, while the billable hour may survive and the associate will adapt, the harder question, the one that should keep the legal industry awake at night, is whether the bench will hold.

You can find more of听our coverage of Legalweek events听here

Key takeaways:

• • • Structural change in firms 鈥� The traditional law firm pyramid, in which junior lawyers perform high-volume work at billable rates, is losing its foundation as AI compresses tasks that once took hours and clients increasingly bring more work in-house.

    • Finding new ways to train 鈥� AI-powered simulations are emerging as a concrete answer to the associate training problem, allowing new lawyers to build courtroom skills faster and fail safely behind closed doors.

    • The associate role isn’t dying, it’s being redefined 鈥� Those law firms that figure out the right mix of legal training, technological fluency, and management skills will have a significant edge over those that are still debating it.

NEW YORK 鈥斕齇n more than one occasion, I have written seriously and at length about the death of the billable hour. I’ve argued that alternative fee arrangements (AFAs) are the future, that the economic logic of hourly billing is irreconcilable with AI-driven productivity gains, and that the industry needs to prepare for a fundamentally different pricing model. I meant every word. I still do.

Yet, at last week鈥檚 one attendee pointed out they鈥檝e been hearing about the death of billable hour since the 1990s. At this point, it’s less a prediction and more of a tradition. Indeed, Matthew Kohel, a partner at Saul Ewing, said despite the legal press coverage connecting AI to the billable hour’s demise that narrative is now entering its third or fourth decade. And Kohel said his firm simply isn’t seeing meaningful client-driven movement toward AFAs.

So let鈥檚 be honest: the billable hour is not dead, and in fact, it may not be even close to dead.

However, if you’re looking for something that is facing a genuine existential reckoning 鈥� something the legal industry whispered about in the early days of generative AI (GenAI) and is now discussing openly 鈥� Legalweek 2026 may have found it. It turns out the billable hour was never the thing in danger, rather it鈥檚 the person billing the hours.

It’s the associate.

The question nobody wanted to ask out loud

The future of the junior lawyer surfaced in virtually every breakout session across the three-days event, and while it may not be the point of inception for the question, it was certainly the moment this idea graduated from a half-whispered aside to main-stage conversation.

Moreover, the problem has grown more urgent since its inception in the early GenAI days, when the question was simply whether a firm would need fewer associates. Now, that question hasn’t gone away, but it’s been joined by harder ones concerning training, hiring, and legal and technical skills. For example, what if AI is already better than a junior associate at some of the tasks that defined the role in the past? And what happens if someone says it out loud?

Someone said it out loud.

If you’re looking for something that is facing a genuine existential reckoning, Legalweek 2026 may have found it. It turns out the billable hour was never the thing in danger, rather it鈥檚 the person billing the hours.听It’s the associate.

During a panel on Measuring What Matters, the conversation turned to client trust. Clients want to know: How can you be sure AI will catch everything? How do you trust it to find what matters across 5,000 pages of documents?

The response from the panel was direct, and it landed like a brick in the room: it’s 5,000 pages, and someone was reading those five thousand pages. That someone is an associate. If that associate 鈥� who, more often than not, is one of the least experienced lawyers in the building 鈥� is the one reading all those pages, why would you trust them to do it better than a machine?

While that question hung in the air during the panel, it does deserve to sit with you for a moment afterward. Because embedded in it is the uncomfortable arithmetic that drives the entire associate question. The traditional law firm pyramid is built on a base of junior lawyers performing high-volume, lower-complexity work such as document review, due diligence, first-pass research, and doing so at rates that generate revenue while the activity is simultaneously (in theory) training the next generation of partners. If AI can do that base-layer work faster, cheaper, and with accuracy that one panelist described as “beyond very good,” then the pyramid doesn’t just shrink. It loses its foundation.

Barclay Blair, Senior Managing Director of AI Innovation at DLA Piper, noted that tasks like due diligence on some types of financial contracts are already being compressed to two hours, down from 15 to 20 鈥� with zero hours being a realistic possibility in the near future.

Further, as one attendee observed, clients increasingly are adopting AI internally, and they’re bringing work in-house that was previously sent to outside counsel. Clearly, the work that trained generations of associates isn’t just being automated 鈥� in some cases, it’s leaving the firm entirely.

Fewer reps, greater weight

Yet here is where it would be easy (and wrong) to write the doom-and-gloom version of the future, in which AI replaces associates, the pipeline collapses, nobody knows how to train lawyers anymore, civilization crumbles, etc. It’s a clean narrative, but it’s also not what Legalweek panels actually said.

Because alongside the anxiety, something else was happening. People were building answers.

In another panel, Developing the Future Lawyer, panelists spent an hour in the weeds of what associate training actually looks like when the old model breaks down 鈥� and the conversation was far more concrete than you might expect.

Panelist spent an hour in the weeds of what associate training actually looks like when the old model breaks down 鈥� and the conversation was far more concrete than you might expect.

Panelist Abdi Shayesteh, Founder and CEO of AltaClaro, laid out the core problem with precision, noting that there’s a growing gap in critical thinking among associates. Templates getting copy-pasted without relevance analysis, and there is a lack of knowing what you don’t know. And the traditional training methods such as videos, lectures, and passive learning, don’t fix it. Indeed, those outdated models may be making it worse. Shayesteh鈥檚 analogy was blunt: You don鈥檛 learn to swim by watching videos 鈥� you need to jump into the deep end.

His solution is AI-powered simulations. Not hypothetical ones, but working deposition simulations available today, with real-time AI feedback, in which associates can practice cross-examination, deal with opposing counsel objections, and build the muscle memory that used to require years of live experience.

Kate Orr, Managing Director of Practice Innovation at Orrick, picked up the thread with two observations that reframed the stakes. First, AI simulations allow associates to fail behind closed doors, a radical improvement over the old model, in which blowing it had real consequences because failure often happened directly in front of the partners Second, the tool isn’t just for juniors. Even experienced lawyers are using simulations to test different approaches, tweak personas, and sharpen arguments. Orrick’s own Supreme Court team had a lawyer use AI to review a draft brief and identify paragraphs that could be tighter.

Todd Heffner, Partner at Smith, Gambrell & Russell, said the real question isn’t whether associates will use AI, but rather whether it gets them to lead at trial in year 10 instead of year 20. Right now, most associates are lucky to see the inside of a courtroom in their first seven years, and even then, they spend most of their time back in the hotel prepping for the more experienced attorneys instead of arguing themselves. If simulations can compress that learning curve, the associate’s career doesn’t disappear, rather, it gets accelerated.

The dinosaur that adapted

During the Measuring What Matters panel, Mitchell Kaplan, Managing Director of Zarwin Baum, introduced himself with a memorable bit of self-deprecation: He’s a dinosaur 鈥� but one, he clarified, who understands how AI can revolutionize what he does.

Kaplan’s perspective threaded through both days of programming like a quiet counterweight to the anxiety. He’d seen this before 鈥� not AI specifically, but the fear of it. He watched the legal industry transition from physical libraries to digital research tools, and he watched attorneys adapt. And his message was consistent: the work changes, but the need for lawyers doesn’t disappear. Associates may be taking shortcuts, but they still need to read, still need to review, and still need to think.

They’re developing differently than his generation did, Kaplan said, but it鈥檚 the same way every generation develops differently from the one before it. And different doesn’t mean wrong.

The work changes, but the need for lawyers doesn’t disappear. Associates may be taking shortcuts, but they still need to read, still need to review, and still need to think.

It’s a perspective that found an unexpected echo in the Enterprise Alignment panel. Mark Brennan, a partner at Hogan Lovells, relayed a comment he heard at a previous AI conference: The next generation of entry-level jobs will be managers 鈥� because they’ll be managing agents and other tech tools. Brennan admitted he didn’t have all the answers on what that means for legal training, but the implication was clear. The associate role isn’t dying, instead, it’s being redefined. And the firms that figure out what that redefined role looks like, what mix of legal training, technological fluency, critical thinking, and management skills it requires, will have a significant advantage over those firms that are still debating it.

Another panelist, Andrew Medeiros, Managing Director of Innovation at Troutman Pepper Locke, made a prediction that felt like the sharpest version of this idea. He said that at some point, new lawyers are going to be doing simulated matters as a standard part of the development process. Eventually, there’s going to be a generation that walks in as new attorneys and finds themselves litigating right away.

That’s not the death of the associate. Rather, that’s the beginning of a different kind of associate 鈥� one who arrives at the courtroom sooner, with different preparation, carrying different tools.

The billable hour, for all the prophecies, refuses to die. The associate, it turns out, has no intention of dying either 鈥� just evolving. Mitchell Kaplan called himself a dinosaur 鈥� but Legalweek was full of dinosaurs, and every one of them was adapting and in that adaptation, thriving. The harder question is whether the firms that forged them will be brave enough to follow.

You can find more of听our coverage of Legalweek events听here

Key insights:

• • • Client expectations around AI have shifted from curiosity to accountability 鈥� Law firms are now being asked not just whether they use GenAI, but to prove how it delivers measurable cost savings on specific matters 鈥� a question most firms still cannot answer with hard data.

    • A growing contradiction defines firm/client relationships 鈥� As clients simultaneously demand AI adoption, require granular billing transparency, and in some cases refuse to pay for work performed with AI, they鈥檙e creating a pricing and value paradox with no clear resolution for their law firms.

    • The ROI challenge around AI is fundamentally a relationship problem 鈥� Driven by a widening gap between what clients expect to save and what firms can demonstrate, a rift has developed between clients and firms, which is compounded by the fact that few firms have a coherent GenAI strategy in place.

NEW YORK 鈥� opened with a keynote conversation featuring Mindy Kaling, the Emmy-nominated writer, producer, and Tony Award-winning playwright, who reflected on a career built around one enduring fascination: messy relationships. She talked about growing up wanting to write something like Sex and the City, only to end up helping to chronicle the internal politics of a Scranton, Pennsylvania paper company in The Office. She talked about her love of watching people navigate breakups and power struggles and then finding the comedy in it all.

If she’s looking for new material, the three standing-room-only panels that followed could keep her busy for seasons.

Not surprisingly, the relationship between clients and their law firms has always been complicated 鈥� bound by mutual need but strained by competing incentives. Now, that tension is starting to reach a rolling boil as many law firms can鈥檛 seem to agree on exactly how the gains of their use of AI tools, especially generative AI (GenAI), are going to be split, or even if they鈥檙e going to be split at all.

AI is no longer optional or experimental 鈥� and many clients simply assume it’s already in use.

Across three 成人VR视频-sponsored sessions during this week鈥檚 Legalweek event, that tension surfaced again and again 鈥� not as a future concern, but as a present reality. Today, clients are arriving at the table more informed, more demanding, and more willing to use AI themselves. Firms are investing heavily in AI, but they still are struggling to quantify returns in terms their clients will accept. With the rates that law firms charge increasing 鈥� averaging more than 7% growth in 2025, and likely to stay on that pace in 2026 鈥� it sets up a collision with savings mandates that have yet to produce a shared framework for measurement. And underneath all of it, a fault line is building pressure 鈥� one that, as Ellen Hudock, GSK’s Chief of Staff Legal and Compliance, is not being resolved.

In 2026, GenAI has become the thing neither side can stop talking about, the thing both sides agree matters, and the thing that neither side can agree on how to handle.

This is not the story of an industry resisting change. Nearly everyone at Legalweek agreed that AI adoption is no longer optional. The harder questions, however, and the ones that echoed through every panel, every audience comment, and every hallway conversation is who benefits, how much, and who gets to decide.

Proving AI鈥檚 path to saving clients money

Three years ago, the client question was simple: Are you using AI, and would you use it on our matters? In 2026, that question has matured, and the new version is much harder to answer.

GSK鈥檚 Hudock described the shift bluntly during one panel. GSK is learning as much as it can from its outside law firms about how they’re deploying GenAI, she said, and are always looking to partner on new use cases. However, she noted that the conversation has moved well past curiosity. The pressure to deliver savings 鈥� internally and externally 鈥� is intense, and the questions have sharpened accordingly: What are you using? How are you using it? How does it generate savings?

Clearly, firms are hearing this message. Matthew Beekhuizen, Chief Pricing and Innovation Officer at Greenberg Traurig, noted that the pace of AI-driven change has accelerated sharply, particularly since October 2025. Clients who had previously said nothing about AI are now asking how it’s being used on their specific legal matters.

Indeed, AI is no longer optional or experimental 鈥� and many clients simply assume it’s already in use, said Mark Brennan, a partner at Hogan Lovells.

The trouble is that firms still can’t give clients the answer they most want to hear. When pressed on how much cost savings AI is actually achieving, the response from the firm side is often: We’re still gathering the data. Mitchell Kaplan, Managing Director of Zarwin Baum, acknowledged the industry is still in the anecdotal phase of measuring returns.

Sergey Polak, Director of Technology Innovation at Ropes & Gray, described the current state of ROI measurement as being based more on conventional wisdom rather than hard evidence. Hudock’s response to this was pointed: That’s exactly the situation in which clients want to partner. Supply the work, and let’s figure it out together.

The contradictions in the room

If the evolution in client expectations were the whole story, it would be manageable; however, the reality is messier than that, because clients are not speaking with one voice.

During another panel, Barclay Blair, Senior Managing Director of AI Innovation at DLA Piper, laid out the contradictions in sharp relief. Blair, who introduced himself as “the extremist on the panel,” is seeing clients who expect AI to be used and are asking how it will achieve specific savings targets. At the same time, many law firms are still receiving directives that feel lifted out of 2023, such as demands for warrants that models are unbiased, and declarations that firms cannot use AI without explicit permission. In 2026, both postures are arriving in the same inbox.

When pressed on how much cost savings AI is actually achieving, the response from the firm side is often: We’re still gathering the data.

The billing conversation captures this tension perfectly. Polak of Ropes & Gray noted that clients are beginning to ask for line-item transparency on invoices 鈥� was AI used on this task, and how much time or money did it save? Simultaneously, as Blair observed, other clients are issuing guidelines stating they won’t pay for certain services if performed by AI. This isn’t clients barring AI outright; rather, its clients demanding firms adopt AI, then using that very adoption as leverage to negotiate a decrease in costs. Not surprisingly, this becomes a self-reinforcing cycle with no obvious exit 鈥� at least, not for law firms.

Meanwhile, Zarwin Baum鈥檚 Kaplan raised a billing paradox that GenAI is making harder to ignore. As AI compresses work that once took hours into minutes, an itemized hourly bill increasingly tells a story that undersells the value delivered. His proposed answer: a return to the single line-item services rendered bill, which actually predated the billable hour. Kaplan then asked whether clients would actually accept it.

The advice to the law firms in the room from DLA Piper鈥檚 Blair was more blunt: Don’t wait for the client to set the terms. Lead the conversation about AI ROI and set the meeting. As Blair described, this is now the time to negotiate how value gets shared, while both sides are still figuring out the rules 鈥� not after one side has already written them.

The pressure hasn’t yet found a release valve

None of these tensions exist in isolation. They are symptoms of a structural mismatch between what clients need from the economics of legal AI and what firms are currently able to demonstrate 鈥� and the numbers suggest the legal industry is less prepared for this conversation than it thinks.

As 成人VR视频’ Steven Petrie pointed out, those law firms with a GenAI strategy are 3.9-times more likely to achieve ROI than those without one. Yet, only 22% of firms have such a strategy, Petrie said. That gap 鈥� between the firms that are thinking systematically about AI’s role in their business and those that aren’t 鈥� may turn out to matter less than the gap between what clients expect to save and what firms can show they’ve delivered.

The ROI question, in other words, is not just a measurement challenge, rather it鈥檚 a relationship challenge. And like all the best relationship drama, the tension doesn’t come from disagreement about whether the relationship matters. It comes from both sides wanting something slightly different from it 鈥� and neither being quite sure if both sides can get what they want.

If Mindy Kaling is still looking for complicated relationships to write about, she knows where to find them. This one鈥檚 going to need a few seasons to work itself out.

You can find more of here

Now scale that up to an entire company. Take pharmaceutical company Organon, for instance. In 2021, Organon was spun off from pharma giant Merck, and suddenly its much smaller legal department had to learn how to handle contracts 鈥� and lots of them. Organon had pre-existing agreements in 37 different countries, all in different formats, and needed to figure out how to simplify and consolidate this hoard of contracts.

鈥淣obody knows where anything is 鈥� for every template we have, there are probably 10 to 15 versions of that template floating around,鈥� recalls Stacy Lettie, Chief of Staff to the General Counsel at Organon. 鈥淭hat in itself creates an inefficiency that is so hard to overcome, it鈥檚 almost a little daunting.鈥�

However, even the greatest challenges can be overcome. At , Lettie and Jamal Brown, Head of Legal Operations and Knowledge Management at JPMorgan Chase, explain how to simplify the complex when it comes to managing the explosion of contracts.

Their takeaway: Such as in life, it takes a village 鈥� and this village includes a combination of people plus technology.

The right tool for the job

Originally, Lettie and the Organon team had a mostly manual process to try and compare and contrast contract templates. At one point, she says, the team took over a whole conference room, printed out as many templates as they could find, and sorted them into piles that could be compared against one another.

Now however, she says that technology provides another option, and it鈥檚 just a matter of finding the right tool for the job. 鈥淲e need to lean into the technology to solve that inefficiency because that is one of the most solvable problems that we have in contract management,鈥� Lettie explains. 鈥淏ut also, it doesn鈥檛 need to be perfect, it doesn鈥檛 need to be a template that solves everything. Let鈥檚 make it good enough.鈥�

Indeed, AI technology is becoming a regular starting point for tasks 鈥� in fact, 82% of corporate C-suites report having used AI as a starting point for tasks, according to data from 成人VR视频鈥� 2024 Future of Professionals Report.

However, not all technology is created equally, the Legalweek panel warns.

Brown says that at JPMorgan, for example, the team has experimented with two separate AI tools for contracting 鈥� and gotten two very different results. The first he called 鈥渁 Cadillac, it was best-in-class and every feature and functionality.鈥� However, it provided a number of solutions to problems the department didn鈥檛 necessarily have. In response, the legal department decided to develop 鈥渁 smaller, medium-value solution that does one thing really well.鈥� And because this solution attacks a single problem, it has been a better value.

鈥淢y recommendation is, don鈥檛 boil the ocean in the first instance that you build,鈥� Brown notes.

The people side of contract tech

With so many different types of contracts to deal with, however, technology is not the only consideration. Brown and Lettie also discussed how to balance standardization and customization 鈥� and importantly, how to make attorneys feel empowered to prioritize what鈥檚 important.

Lettie notes that at Organon, the legal department does not actually own the contracting processes, the business side does. The legal team gives the templates and the playbook, but at times, those templates are not always followed, and the business side accepts the client side鈥檚 contract as the basis of the agreement.

What results is not a technological question, but a business one. 鈥淚 felt that our younger lawyers in particular didn鈥檛 feel empowered,鈥� Lettie says. 鈥淭hey had no basis to say, 鈥楴o, I鈥檓 not going to review that.鈥欌��

“We need to lean into the technology to solve that inefficiency because that is one of the most solvable problems that we have in contract management; but also, it doesn鈥檛 need to be perfect, it doesn鈥檛 need to be a template that solves everything. Let鈥檚 make it good enough.鈥�

In this case, she explains, tying contracting decisions to the business at large has helped her adopt a strong stance in dealing with the business. 鈥淭here are certainly things you need to guard against, but honestly, if your non-disclosure is eight years versus three years, who knows, who cares?鈥� she says, adding that it鈥檚 not worth an attorney鈥檚 time when they need to provide value to the business. 鈥淩eviewing an NDA is not any value to anybody.鈥�

And it鈥檚 in these kind of human-centric decisions in which technology can play a valued-added role, especially as this technology continues to evolve. 鈥淭here is no reason to even really be having that my paper/your paper discussion,鈥� Lettie notes. Today, contract technology can take a template, turn it into a playbook, and put it against that third-party contract. 鈥淚t comes in, it goes in the engine, you get a comparison, you sign it, or you don鈥檛, and you move on with your day. It shouldn鈥檛 take longer than 30 minutes.鈥�

Brown agrees, noting that in evaluating solutions, he comes back to the question: 鈥淲hat do I want my lawyers to be working on? Working on a multi-million-dollar M&A deal, or working on a single paper?鈥�

To help free up that time, JPMorgan Chase鈥檚 legal team has developed a suite of 11 GenAI models, most slated for knowledge management, but that can be used across entire product line. This process provides a host of new capabilities, such as the ability to ask direct questions about contracts and documents. 鈥淲e built models that have data across all of our products and services globally,鈥� Brown adds. 鈥淚t makes for a more intelligent way for the solution to interact with our internal professionals.鈥�

That scale of technology-build may not be right for every legal department; but both Lettie and Brown agree that legal departments should be thinking about not just what the technology can do, but how it fits into the overall collaborative team picture. Departments also need to examine their ability to accept failure if it does not work.

Brown tells a story from two years ago, about a contract vendor that had a fantastic pitch to solve a crucial problem, However, the in-house trial wasn鈥檛 going well. Rather than push forward unnecessarily, the in-house team decided to take a step back. 鈥淭hank god we did that,鈥� Brown says, 鈥渂ecause we were able to recover and prepare for the next wave, which was [Chat]GPT.鈥�

The result is a lesson: The whole team needs to be on board to truly innovate. 鈥淪tart small and fail fast,鈥� Brown says today. 鈥淒o not be afraid to let leadership know that something鈥檚 not going right.鈥�

You can find more听coverage of LegalWeek events听over the years here

However, there is a difference between simply throwing money at technology and fully integrating it into a firm鈥檚 strategic plan. To hear panelists at event tell it, firms are fully aware of the difference. 鈥淭ech investments are and always should be increasing,鈥� says Elaine Dick, Knowledge Management & Data Services Manager at BakerHostetler. 鈥淏ut what鈥檚 a bit more important with that increase is having it adapt with what鈥檚 happening with technology.鈥�

The key comes in finding how to manage those budgets in a rapidly advancing technological world.

The trade-offs for new tech investment

Don Sternfeld, Chief Innovation Officer at Steptoe, joked that his firm chair 鈥渉as never walked in and said 鈥楽ternfeld, you鈥檝e got an unlimited budget鈥欌�� 鈥� but that鈥檚 the common dilemma at today鈥檚 firms. 鈥淭he challenge that we have is, if we want to make those significant investments and meet firm goals for profitability, how do we minimize or at least contain the ones that are required to keep our lights on?鈥�

Sternfeld says that since arriving at the firm, he鈥檚 aimed to engage in zero-sum budgeting 鈥� making sure that income minus expenses always equals zero. 鈥淥ur role is, at the end of the day, to maximize firm and partner profitability,鈥� he explains. 鈥淲here are the tradeoffs?鈥� There is no usual place, he adds, but in practice a firm tech leader may find them within other technologies or other functional areas, or by simply finding more efficient ways to operate within the organization.

Kathleen Orr, Global Head of Practice Innovation at Orrick, Herrington & Sutcliffe, adds that she aims to achieve this budget balance by directly tying technology priorities to strategic areas of the firm. Or simply put another way, if there鈥檚 a practice area that鈥檚 growing, they probably are a better candidate for technology investment. It seems simple, she notes, but some firms do not keep an eye on growth when determining tech investment. 鈥淚 would underscore how important it is to understand those priorities, to have those conversations,鈥� Orr says.

Maximizing attorney engagement is key to getting the most out of more expensive tech investments, and that gives small and midsize law firms the ability to compete across the board because of advanced tech tools and services.

At Orrick, Orr says that she has a set of standard questions that allows apples-to-apples comparisons of tech priorities: What is the use case for this? Who has the use case for this? Does the firm already have something that does this? 鈥淪o, it鈥檚 not just up to me to check the box and say, 鈥榊up, that looks great!鈥欌�� she explains. 鈥淲e have a process.鈥�

Sometimes, however, a more hands-on approach is preferred. Dick at BakerHostetler says her team has instituted the cleverly named Bake-Offs by putting two pieces of technology head-to-head in limited trials for a select group of interested attorneys. In this way, rather than relying on a vendor鈥檚 perception of how a technology would work, the firm鈥檚 attorneys can provide feedback directly 鈥� before the firm makes a large-scale investment.

鈥淚f it doesn鈥檛 work in an attorney鈥檚 process, there鈥檚 no point in buying it,鈥� Dick explains. 鈥淏ake-Offs lets the ROI be driven by the business or the attorneys that would be utilizing it, or the attorney that specifically asked for it.鈥�

The people behind the metrics

BakerHostetler鈥檚 Bake-Offs also illustrate one of the panel鈥檚 crucial points: Not all metrics merely lead back to dollars and cents. 鈥淎ttorneys are artists, not machines,鈥� notes Martin Kilmer, Chief Operating Officer at Gravis Law. 鈥淵ou have to convince them that it will benefit the product that they are creating.鈥�

To that end, when analyzing technology metrics, one of the first measurements that Kilmer turns to is user adoption. 鈥淵our development road map has to address pain points,鈥� he explains. 鈥淭he metric you have to utilize is adoption, and if you鈥檙e not getting that, it will fall flat on its face.鈥�

This is also why Kilmer said his firm commits to ongoing training. Especially at a smaller firm, maximizing attorney engagement is key to getting the most out of more expensive tech investments, he says, adding that small and midsize law firms now have the ability to compete across the board because of advanced tech tools and services.

However, that means that the tech function needs to have regular conversations with attorneys to determine what they actually need and how those needs may change over time. 鈥淚鈥檓 convinced that three-quarters of my attorneys have no idea the power they have in their hands right now,鈥� Kilmer adds.

Similarly, Orr said she runs into three common pitfalls with tech adoption. First, is that tech is an easy sell鈥� but 鈥測ou need people that were involved in the selection of the tool to help you sell it.鈥� Second, is that if you build it, they will come鈥� but 鈥測ou have to tell people things many, many times, and you have to tell them how it will benefit them.鈥� And third, is that with tech, you can fix it and forget it.

鈥淵ou really need to be thoughtful at the start about how you鈥檙e going to define success,鈥� Orr adds.

The ultimate judges of success

Ultimately, as the attorneys are the end arbiters of this technology, they will largely judge its success or failure. That means getting attorneys on board with tech adoption from the beginning and integrating them into the metrics collection process is crucial. And while this can be easier said than done, having attorney champions can help, Sternfeld says.

鈥淭hey need to know, 鈥榃hat鈥檚 in it for me?鈥� It鈥檚 all about organizational change management,鈥� he explains. 鈥淎nd the best people to tell them what鈥檚 in it for them, is that other lawyer. They won鈥檛 listen to my emails, but they鈥檒l listen to their peer in that meeting who says, 鈥極h, this is something out there.鈥欌��

By providing the tech platform and the metrics to prove its worth, the tech function and attorneys can become partners in the success of any tech adoption. 鈥淎t the end of the day, the innovation that our clients really want is brilliant legal practice,鈥� Sternfeld says, adding that the tech function鈥檚 role in law firms will become 鈥渉ow can you give tools that let [lawyers] be the best and get rid of the noise.鈥�

You can find more coverage of LegalWeek events over the years here

However, particularly in legal, there remains two major sticking points to widespread adoption: ethics and regulation. While the technology itself is rapidly emerging, Gen AI is not likely to see widespread adoption in the legal industry if it can’t be trusted, and many legal practitioners are waiting for additional guidance before using Gen AI to tackle more widespread tasks.

According to a panel, The Ethical and Regulatory Impacts of Using Your Data to Train AI, at the听, that guidance is coming in the United States 鈥� albeit slowly, and in parts. However, the guidance that is available now is enough to get started on Gen AI risk mitigation planning, panelists said.

US Gen AI regulation today

Across most US jurisdictions, Gen AI regulation is extremely active but is still in the early stages. For example, the panelists noted that many US states are beginning to form councils and task forces to look into AI. 鈥淲e鈥檙e now seeing this over-arching focus on the risks鈥� but that trickles down into the state legislatures themselves,鈥� said panelist Garylene Javier, a Privacy & Cybersecurity Associate at Crowell & Moring.

In particular, a lot of the AI focus has been around the privacy rights of individuals, particularly around consumer protection and the right to opt out of AI systems. Javier explained that all states (except Utah) that have adopted privacy laws have done so with opt-out provisions as part of the legislation. 鈥淲hat we鈥檙e seeing is a shift of all of these different privacy laws with that particular focus, particularly when it comes to automated decision-making,鈥� she explained.

Anyone in this space needs to be thinking through the scenarios.

On a federal level, President Biden鈥檚 2023 executive order on AI remains the most relevant guidance, particularly as the in early February that all executive agencies have completed the 90-day actions tasked by the order. Panelist Ignatius Grande, Director at Berkeley Research Group, cited the Federal Trade Commission (FTC) as one regulatory agency to watch, explaining it has been among the most proactive in the AI space.

鈥淭hey鈥檙e coming out and saying, there are copyright issues with LLMs [large language models] and AI,鈥� Grande said.

Of course, the upcoming 2024 presidential election could cause upheaval for AI regulation. As for exactly how, the panel did not speculate 鈥� only to say that it could have wide-reaching ramifications. Jason Winmill, Chair of legal industry organization Buying Legal Council, observed that the previous two presidential administrations viewed technology regulation very differently, and 鈥渁nyone in this space needs to be thinking through the scenarios.鈥�

Winmill drew an analogy to a pool table, where the pool balls are not independent, but interact with one another in seemingly endless varieties. What matters, however, is who鈥檚 lining up the shot. 鈥淲e don鈥檛 know what that pool table is going to look like in 2 to 3 years.鈥�

The legal industry鈥檚 response

Given the potentially transformative nature of Gen AI, it鈥檚 reasonable to expect that it will see widespread usage within the legal industry within the next 5 to 10 years; but the panel noted that in order to do Gen AI right, corporate legal departments and law firms will need to do their due diligence.

Grande told the story of Samsung, which ran into a problem last year with employees using ChatGPT. One employee uploaded source code to ChatGPT that was viewed externally, while another employee put a confidential recording into the tool to create a summary. Samsung ended up until it could put more controlled Gen AI systems in place.

There鈥檚 a lesson for the legal industry no matter where you sit, Grande added. 鈥淎s far as what law firms can be doing, part one is understanding and giving guidance to their employees鈥� and have in place acceptable ways of [using] generative AI. There are a lot of questions that you need to ask, and if you鈥檙e someone who鈥檚 not experienced in the area, it鈥檚 really hard to know what questions to ask.鈥�

Winmill agreed, noting that from the in-house legal perspective, AI use needs to be ethical but also should be 鈥渄riven by real business needs.鈥� Answering the relevant AI questions starts with getting a number of different parties in the room: legal, for certain, but also procurement, IT experts, AI experts, and business advisors. Concerning the ethics risks, he added: 鈥淭he space is just moving way too fast, and you鈥檙e going to have to have outside counsel or advisors who can advise on a timely basis.鈥�

There are a lot of questions that you need to ask, and if you鈥檙e someone who鈥檚 not experienced in the area, it鈥檚 really hard to know what questions to ask.

Of course, that doesn鈥檛 mean throwing as many people as possible at the AI problem. The key is to be strategic, Winmill said, and bigger teams aren鈥檛 always better. 鈥淭he size of the table is growing, it鈥檚 growing faster, and that is also a concern,鈥� he explained, noting that academic literature shows larger teams can have communication issues. 鈥淲e鈥檙e adding more resources in, but we鈥檙e also going to be encountering more problems.鈥�

Communication is also a key when outside forces are involved, particularly when firms are dealing with outside regulators. Corporate legal departments and law firms need to demonstrate first that they have a plan, Javier said, and then demonstrate that they are doing everything they can to stick to that plan. This has been a particular point of emphasis when analyzing AI for potential bias, which has increasingly been the subject of US state legislation and lawsuits.

Organizations should be 鈥渕aking sure from an ethical standpoint, you鈥檙e demonstrating across the entire process [that] you did as an organization or a law firm take the time to step through and see how it affects certain demographics or customers,鈥� Javier said, adding that as hard as legal departments or law firms may try, any Gen AI policy is 鈥渘ot going to 100%鈥� in terms of catching every problem.

鈥淪omething will happen tomorrow where it鈥檚 going to be totally deficient,鈥� she said. 鈥淭he best thing we can do is try 鈥� try really hard.鈥�

How can lawyers and IT personnel keep up with the cyber-threat onslaught? It starts with a simple mantra: Nail the basics.

At the Navigating the Cyber Threat Terrain: Cybersecurity, Privacy and Legal Sector Focus panel during this week in New York City, cyber-attorneys and experts from companies and law firms assembled to give their advice and experience on how to keep up with emerging threats.

Always aware of everything

One of the biggest challenges, the panel noted, is simply staying aware of the mass of cybersecurity and privacy rules and regulations, particularly for organizations that operate on a global scale. Panel moderator Manny Sahota, Director for Global Cloud Privacy, Regulatory Risk & Compliance at Microsoft, noted that while everyone may have focused on rules coming out of the EU and US recently, simultaneously, Chile updated its security regulations for the first time since 1999.

Even once the legal and IT teams are able to understand the situation, however, there remains the issue of getting others in the organization to care.

It’s a lot to follow but also next to impossible to predict, agreed Daniel Ostrach, Senior Corporate Counsel at Microsoft. 鈥淥ne of the hardest things for us to do is anticipate the way that regulators are thinking 鈥� but we can鈥檛 run our business based on yesterday鈥檚 regulation,鈥� he explained. However, in today鈥檚 climate, just following the regulation 鈥渋s the bare minimum, that鈥檚 table stakes.鈥�

Sabrina Ceccarelli, Global Vice President and Assistant General Counsel of Commercial at Lightspeed Commerce, gave the example of one recent privacy regulation: Quebec鈥檚 Law25, which is more similar to the EU鈥檚 General Data Protection Regulation (GDPR) than other Canadian privacy laws. Without enough privacy staff to keep up, her team turned to the privacy resources they did have: 鈥淲e do as much rinse and repeat as we can.鈥� They looked at areas such as training in which they already had pre-established guidance, then updated rather than reinventing the wheel.

Even once the legal and IT teams are able to understand the situation, however, there remains the issue of getting others in the organization to care. Joseph Lee, Director for Information Security & Compliance at law firm Arnold & Porter, said that his most effective method is simple: 鈥淏ombard people over and over and over.鈥� Constant reminders and messaging from multiple sources such as town halls helps people realize that cybersecurity is not a set-it-and-forget-it proposition, Lee said. 鈥淚f you just do an annual training, it鈥檚 not bad, you check a box, but that doesn鈥檛 keep it top of mind.鈥�

From the technology standpoint, Rachi Messing, Co-Founder of startup Altorney, also noted that legal has an opportunity to work with engineering to make sure privacy and security is evident in everything they do. For instance, Messing noted that every development ticket or feature request at the company has a mandatory security and privacy analysis. That analysis is 鈥渘ot just a check box,鈥� he said, but forces tech teams to think through potential impacts and why they occur. 鈥淭hat really does force a focus in the culture of, How are we focusing on security? How are we focusing on privacy in everything that we do? Otherwise, that鈥檚 how you find yourself on the front page of The New York Times.鈥�

Cyber Dungeons & Dragons

Once the awareness has been achieved, then it falls on the legal, IT, and other security and privacy-related teams to execute. Once upon a time, those teams might have all been separate entities, the panel noted, but Messing added: 鈥淭he truth is, in today鈥檚 world, there really can鈥檛 be a gap.鈥�

At his startup, Messing said he and his co-founders did not have the ability for a formal chief information security officer (CISO) or privacy team. However, they picked outside counsel based explicitly on the firm鈥檚 ability to support the company around security, advise on privacy, and then work with the company鈥檚 engineers. 鈥淲orking together there is the only way that a company is going to be able to succeed,鈥� Messing explained. 鈥淚f the two sides are feuding with one another鈥� you鈥檙e never going to be able to survive in today鈥檚 world.鈥�

Lightspeed鈥檚 Ceccarelli agreed, noting that the role of the corporate lawyer has changed. She says her legal team鈥檚 mantra last year was 鈥�We鈥檙e building GCs,鈥� noting that for many corporate attorneys, the GC chair is their ultimate goal. However, implicit in that is that 鈥渘one of us can call ourselves an excellent tech lawyer if we don鈥檛 understand privacy.鈥� As a result, her team created knowledge-sharing exercises with continuous updates, which created some ownership and accountability for the legal department to work with the whole enterprise. 鈥淟egal counsel can鈥檛 just be doing contracts anymore,鈥� she said. 鈥淲e need to be more than that.鈥�

The panel cautioned to make sure that not only is everybody speaking to one another 鈥� especially the lawyers 鈥� but they are speaking the same language when making these plans.

One way to make sure the organization comes together is through tabletop exercise, the panel suggested. Lee admitted that 鈥渢he tabletop exercise may seem like a corporate Dungeons & Dragons sort of thing,鈥� but added that it鈥檚 really important to go through potential risky scenarios. 鈥淚f you don鈥檛 have a plan of action, I make an analogy like it鈥檚 a kids鈥� soccer game, everybody is just going towards the ball,鈥� he explained. Tabletop exercise helps answer some basic questions: Who鈥檚 doing negotiations? Who鈥檚 going to the insurance carrier? Who鈥檚 doing communications, and how much?

From there, Ceccarelli suggested making a formal playbook, to make the process memorable and repeatable. The playbook should include engineering and IT, certainly, but it also gives the legal team a seat at the table to help guard against risk and potential worst-case scenarios. 鈥淏y doing that, you can proceed rather quickly but also mitigating any possible damages from the incident that has occurred,鈥� she added.

Finally, the panel cautioned to make sure that not only is everybody speaking to one another 鈥� especially the lawyers 鈥� but they are speaking the same language when making these plans. Microsoft鈥檚 Ostrach gave the example of a three-page legal memo that might give all of the relevant information on a new regulation but would never be read by engineers 鈥渟o it鈥檚 worthless.鈥� In addition to being a lawyer, today鈥檚 counsel need to be 鈥渁n old-timey phone connector,鈥� making sure that everybody is communicating with one another.

And that goes both ways, Lee of Arnold & Porter added. 鈥淚f you鈥檙e in IT and you鈥檙e not regularly talking to your general counsel, you should.鈥� Perhaps the best thing that all parties can do when it comes to privacy and security is a simple trick, he added: 鈥淏e proactive in terms of having those conversations.鈥�

In fact, law firms that use ALSPs are doing so for more than 40% of their intellectual property management work and more than 20% of their legal drafting and legal research, according to the 成人VR视频 Institute鈥檚 recently published ALSP 2023 Report. Corporate law departments that use ALSPs, meanwhile, are tapping them for more than one-third of their regulatory risk & compliance matters and contract management tasks.

Yet, despite ALSPs鈥� rising prominence, the report also reveals that law firms and corporate law departments alike don鈥檛 yet fully trust ALSPs. Almost two-thirds (62%) of law firms indicated that concerns about quality affects their willingness to use ALSPs; and 46% of corporate clients said the same. Meanwhile, more than half of law firms had concerns about turning over confidential client information to ALSPs, and firms also felt their traditional business model was challenged by ALSPs鈥� use of technology. Further, 38% of corporate law departments indicated they would rather have their outside law firms deal with ALSPs rather than deal with ALSPs directly themselves.

So where is the disconnect? At a recent听听session, Changing Nature of Legal Practice: Impact of ALSPs, Tech Companies, and the Big 4, panelists explored not only how ALSPs are increasingly entering legal matters, but what they will need to do to keep their place at the table. No surprisingly, it all starts with the relationship, said panelist Vedika Mehera, Director of Orrick Labs at Orrick, Herrington & Sutcliffe. 鈥淲e can鈥檛 overemphasize how important trust is, and it begins with constant communication,鈥� Mehera said, adding that many view this as constant communication of problems.

鈥淲e don鈥檛 get to fail more than once [in the legal industry],鈥� she explained, noting that there is a flip side to that 鈥� proactively explaining how ALSPs can help. 鈥淲hen you have a success, share it. When you have a challenge that you鈥檙e running up against, communicate it. 鈥 think that really helps build trust.鈥�

Showing clients the value

Of course, this can mean communicating positive return-on-investment for a matter as well. However, as panelist David O’Hara, Director of Legal Business Solutions at Big 4 firm PwC, added: 鈥淚t鈥檚 not just about the dollars and cents.鈥�

O’Hara noted that PwC has worked to bring in diverse teams to matters, in some cases pairing legal experts with IT, bringing in regulatory experts into a cross-border matter, or even identifying internal skills 鈥渢hat helps us build trust with clients, to say we have a world of resources鈥� to tackle different jobs.

鈥淭he more [we] can ease that [concern] and create different ways of working for them, that鈥檚 the intangible value we always need to remember,鈥� he noted.

Of course, this can be easier said than done. Often, clients may not even know the different jobs that an ALSP can do. Indeed, 33% of global corporations within the ALSP Report said that not being aware of services or where to find them was a factor in not using ALSPs.

To help solve for this, O鈥橦ara recommended approaching the problem with more than just a technology solution. 鈥淎ttorneys are going to forever be skeptical of technology, and that鈥檚 good,鈥� he explained, adding that by centering the human and casting the ALSP鈥檚 offerings as part of a wider team鈥檚 efforts, ALSPs can help ease client fears of trying something new.

鈥淲e can鈥檛 overemphasize how important trust is, and it begins with constant communication.鈥�

鈥淎s we integrate all of these different elements, that鈥檚 where I鈥檝e seen the most success,鈥� O鈥橦ara said. 鈥淚鈥檝e seen that be the best delivery of the best combination and use case for how to leverage these teams.鈥�

Mehera added that identifying the client鈥檚 culture is important when approaching those conversations 鈥� as is reflecting how a team with diverse skills can supplement that culture. For example, some corporate law departments may be more tech-savvy than others. 鈥淎ll of these clients are using AI in their business. Can we be reflective of that as well?鈥� Mehera asked.

Finally, O鈥橦ara noted that it鈥檚 important to be realistic when talking about what an ALSP can do. 鈥淲e realize that driving a lot of these changes takes an investment. We鈥檙e pragmatic鈥� and not guaranteeing immediate return-on-investment if it鈥檚 a long-term project, he explained.

O鈥橦ara also said that rather than looking for new tools, lawyers often take the mentality that 鈥渨e need more hammers to break more rocks.鈥� However, for those that take the plunge to try a new path, ALSPs can supercharge a matter 鈥� a value that law firms and corporate law departments alike are beginning to see. More than two-thirds of law firms (69%), in fact, believed that using ALSPs can help them scale and expand their own business.

鈥淚 think there鈥檚 all sorts of different ways that you can partner with vendors and providers in the industry to accomplish what you need to,鈥� Mehera added. 鈥淚 think it鈥檚 in our best interest that everyone succeeds. If you succeed, we succeed.鈥�

However, the reality remains far from simple. There are layers of algorithms that create content (thus the generative part of the name) by continually predicting the next word, with extremely large data sets needed to make these predictions accurate. Further, the newly-released GPT-4 is even multi-modal, meaning it can accept both text and image inputs, which ratchets up the complexity even further.

As a result, even among the most optimistic technologists, there remains some generative AI risks that can鈥檛 be ignored. And as a key panel, Reshaping the Legal Profession: Thriving in the Age of Generative AI & ChatGPT, at the recent explored, the heavily hyped technology may be less of a do-it-all tool and more of a 鈥渕oderately bright, but very lazy first-year associate.鈥�

What generative AI is & what it isn鈥檛

That comparison came courtesy of panelist Aaron Crews, currently Chief Product & Innovation Officer at alternative legal service provider UnitedLex and formerly Chief Data Analytics Officer at law firm Littler Mendelson. Crews noted that while many legaltech types have high hopes for generative AI use in law, including himself, at its core the technology isn鈥檛 that revolutionary.

鈥淕enerative AI is fancy marketing-speak for a machine that anticipates where you want to go next,鈥� he said, adding that while there may be high expectations of a tool named artificial intelligence, in reality 鈥渋t鈥檚 not intelligent.鈥�

Indeed, generative AI is bounded by the data that is put into the system. That means that ChatGPT, developed by OpenAI and currently the most famous generative AI platform, has access to untold amounts of data to make its predictions 鈥� but that data is only current as of 2021, meaning it cannot adjust to newer events.

The tool also suffers from 鈥渉allucinations,鈥� meaning that sometimes the technology 鈥減redicts鈥� facts that have no actual basis in reality. In one notable case, as explained by panelist Foster Sayers, General Counsel & Chief Evangelist at software company Pramata, a Michigan judge tried asking ChatGPT about why a certain court decision was decided the way it was and found that ChatGPT completely made-up precedential cases 鈥� something the judge caught easily, since he had decided the case himself.

With the recent release of GPT-4, a factual accuracy rate between 70% and 80%, depending on the subject matter. But that 20% below perfect is 鈥渟ignificant鈥� in law, explained another panelist, Ilona Logvinova, Associate General Counsel at McKinsey & Company. Technologists in law often run into risk-averse attorneys and clients, where one bad experience can lead to a closed door for all future technological advancements. And although some companies and are hiring for a new role known as a prompt engineer to ask generative AI platforms more specific questions to get a desired outcome, it鈥檚 impossible to create a foolproof system.

鈥淧rompt engineers are getting more popular, but they鈥檙e also learning on the spot,鈥� Logvinova noted.

So where鈥檚 the use?

That鈥檚 not to say that generative AI will fall by the wayside, however. The panel identified a few potential use cases for generative AI in professional services as it now stands: document analysis, review and drafting; research and knowledge management; contract analysis and drafting; and chatbots and assistants. However, the technology is moving quickly, and so too are its potential applications, panelists added.

One panelist, Danielle Benecke, Founder of Baker McKenzie Machine Learning at law firm Baker McKenzie, noted that 鈥渇irms and other enterprises have been sitting on this unstructured data forever,鈥� which generative AI can help unlock.

However, while many regular generative AI use cases focus on the wide data sets the tool already has, it鈥檚 more interesting to start with the enterprise鈥檚 data and running AI against it, Benecke explained, adding that, for example, a firm鈥檚 M&A deal room could run generative AI against the data set and theoretically create a due diligence checklist based on the firm鈥檚 contracts that are already in place.

Pramata鈥檚 Sayers did question how much better generative AI is at producing new documents and contracts than simply using regular templates. While generative AI may produce bespoke work product, legal documents often have to be worded in a very specific way that鈥檚 tough to predict, he said. Contract experts such as TermScout鈥檚 Evan Harris have , finding that while generative AI can create a passable first draft contract, the outputs still require a good deal of editing and governance.

With these limitations in mind, Logvinova added 鈥渋t鈥檚 safer and less risky鈥� to use generative AI for internal purposes rather than for client-facing content or communications. Crews agreed, saying that he 鈥渁bsolutely would not鈥� use ChatGPT for client work as the technology currently stands, but that it may be helpful in fast-forwarding the data creation and ingestion process.

No matter the use case, however, all panelists agreed it鈥檚 paramount to avoid the temptation of adopting generative AI just out of curiosity. Due to its risks, and with the technology in its early stages, any use should be conducted with the firm鈥檚 overall data strategy underpinning the AI use and with a specific goal in mind.

Benecke said her exploration of generative AI primarily focuses on holistic applications across the firm rather than one-off use cases. Any time the firm adopts an AI tool, she said, it鈥檚 with the specific goal to 鈥渟upercharge the firm鈥檚 most valuable pre-existing service lines,鈥� directly tying the AI use with a firm strategic initiative.

Still, there remains a number of unknowns about generative AI鈥檚 use in professional services, and the balance between risk and innovation with generative AI weighing on the scale is one that firms are still working out.

鈥淏e forward-leaning, but be smart about your governance,鈥� Benecke warned. 鈥淵ou don鈥檛 want to be that cautionary tale.鈥�