Summary

In episode 24 of 2030 Vision, hosts Jen Leonard and Bridget McCormack explore the rapidly evolving intersection of artificial intelligence and legal practice. From personal experiences with AI tools to regulatory breakthroughs in the U.K., the episode weaves together seemingly disparate developments into a powerful narrative: the legal profession is on the brink of fundamental transformation.

Highlights include OpenAI's leadership reshuffle, Google’s reckoning with the rise of reasoning models like ChatGPT, and the groundbreaking approval of Garfield Law—the first AI-native law firm authorized to deliver legal services independently. The hosts also discuss how legal regulators in the U.S. may struggle to keep up and how this technological leap could outpace the paraprofessional movement entirely.

Key Takeaways

1. AI Is Becoming a Daily Tool—Even for Mushrooms

Co-host Bridget McCormack recounted using ChatGPT to identify mushrooms during a hike and to prepare for a live TV segment. Her experience shows how AI is not just reshaping high-level legal workflows—it’s becoming part of everyday life.

2. ChatGPT Can Be a Brand Consultant

Jen Leonard used ChatGPT and the Ogilvy taxonomy to evaluate her website, receiving detailed and actionable feedback. The lesson? AI is a creative collaborator, even for those outside traditional tech or marketing sectors.

3. OpenAI Restructures—And the Implications Are Big

With the appointment of Fidji Simo as CEO of Applications, OpenAI is signaling a shift toward productization and commercialization. Sam Altman’s continued focus on AGI (Artificial General Intelligence) suggests long-term strategy divergence.

4. Google’s Search Empire Is Under Siege

The DOJ’s antitrust trial revealed that Apple users are turning to AI models like ChatGPT and Perplexity instead of Google Search. As reasoning models become more accurate and useful, traditional search may become obsolete.

5. The U.K.’s Garfield Law Sets a Global Precedent

Garfield Law (also known as Garfield AI) was approved to deliver legal services entirely through AI. Designed for small claims and debt recovery, it represents a tectonic shift: not just “lawyers using AI,” but AI as the lawyer. The use case is narrow, but the implications are vast.

6. U.S. Regulation May Be Too Slow to Respond

American legal regulators are still piloting paraprofessional programs while the U.K. moves full-speed into AI-enabled service delivery. With 50 state-specific jurisdictions, the U.S. regulatory model may be too fragmented to adapt quickly.

7. “Bionic Boutiques” Could Disrupt Big Law

Inspired by Jae Um’s reporting, the hosts explore the idea of elite partners leaving Big Law to form tech-enabled “Bionic Boutiques.” These firms could use AI agents to handle operational tasks while the attorneys focus on high-level strategy and client service.

8. Regulation Should Balance Risk and Access

In the U.K., legal regulation is framed as both a way to protect the public and to expand access to justice. In the U.S., the focus tends to lean heavily toward the former. This difference in philosophy could widen the innovation gap.

9. We May Be Watching Legal Disruption from the Sidelines

As tech companies like OpenAI move further into commercial spaces, the legal field risks being reshaped by outsiders. If the profession doesn’t proactively innovate, it could be disrupted by default.

Final Thoughts

The episode leaves listeners with an urgent message: AI isn’t just a tool for tomorrow—it’s a force transforming the profession today. Legal professionals who ignore it do so at their peril. With regulators moving slowly and innovation accelerating, the power to shape the future may fall not to lawyers, but to technologists.

As Bridget said: "It is inevitable. The only real question is: Who do you want making decisions about the future of our profession?"

Whether you're a junior associate, a managing partner, or a legal innovator, now is the time to get curious, get engaged, and get building.

Watch Episode 24

Transcript

Jen Leonard: Hi, everybody, and welcome to 2030 Vision: AI and the Future of Law. I’m your co-host, Jen Leonard, founder of Creative Lawyers, here with Bridget McCormack, president and CEO of the American Arbitration Association. And we are, again, in person in Philadelphia. This is very exciting. Hi, Bridget.

Bridget McCormack: Hi, Jen. It’s very exciting. I feel like we’re making a habit of it—and I kind of like it.

Jen Leonard: I kind of like it as well. Seeing you in three dimensions—very nice. I’m happy to be here with you. And we’re going to talk today about a really exciting development, not in Philadelphia, but in the U.K.—one of the first AI-native law firms in the world, probably. But first, we’ll start, as always, with our two segments: AI Aha!, where we talk about something we’ve used AI for recently, and What Just Happened?, where we catch up on recent developments in tech. So I’m going to turn it to you—because you always have the most interesting ahas. What’s yours this week?

AI Aha! Foraging for Mushrooms and Feedback? There’s an App for That

Bridget McCormack: I don’t know if mine are interesting, but they’ve become habits. My husband and I were visiting some friends this weekend who have a home in the Shenandoah Mountains—we took a hike on the Appalachian Trail, which was amazing. I’d never done that before. One of our friends kept texting photos to another friend, who’s great at identifying plants. And I was like—“I have a friend with me right now who can help us.” So we started using ChatGPT to identify mushrooms, which was really fun. We were trying to find edible ones—not all mushrooms are safe, so that was helpful.

And I did one other thing—a little twist on something I’ve done before. I had to do a live TV interview on the CBS Morning Show. It was a totally friendly segment about the benefits of mediation, especially for folks who lost their homes in the fires. Mediation was the focus—not me, not AAA, just mediation. So I asked ChatGPT to practice with me and then give feedback. That was great.

But then I flipped it—I asked it to answer the questions, so I could see the answers it would give. It was a really nice extra way to prepare. Highly recommend that if you’re prepping for something live.

Jen Leonard: Did you feel like anyone actually asked you questions that were helped by your prep?

Bridget McCormack: I think I might’ve been fine even if I hadn’t prepped, but it definitely made me feel more comfortable. I will say—there was another person being interviewed, a mediator from a community center who had lost his home—just a really sad story. They asked how people could get in touch with his center, and he kind of froze. He said, “I think the yellow pages?” I felt bad that I couldn't jump in. ChatGPT did not prepare me for that moment.

Jen Leonard: Come on, ChatGPT!

Bridget McCormack: How about you? What’s your AI Aha! this week?

Jen Leonard: I don’t know if this is a good one, but I thought it was interesting. People are so generous on LinkedIn—they share all these different ways they’re using AI. I don’t know how the algorithm decides what to show me, but someone in marketing showed up on my homepage. They were feeding their company’s website through ChatGPT and asking it to apply something called the Ogilvy taxonomy.

Bridget McCormack: I don’t even know what that is.

Jen Leonard: Same—I had no idea. Another reason I love AI: I don’t really need to know. But apparently, it’s a 15-point scale that marketers use to evaluate the quality of a website. Each item on the scale gets scored between six and seven points. So I took this person’s prompt—thank you to whoever you are—and I fed my own website into ChatGPT. I asked it to score it using the Ogilvy taxonomy and give me a report card. And it did.

Some of the suggestions were irrelevant for my audience—like “ask me now,” or “schedule your free 15-minute consult,” which doesn’t quite fit for me. But others were really good. So I had it revise the suggestions based on my audience—law firms and corporate legal—and it gave me a much more professional version. Honestly, I wish I had known about this before we built the site.

Bridget McCormack: Did you end up making changes as a result?

Jen Leonard: I don’t know how to make changes yet. That’s a whole other podcast—building a website is not something I enjoy. But if I were starting over, I’d begin with the Ogilvy taxonomy, which I assume marketers do.

Bridget McCormack: We’re in the final stages of standing up a new website—we’re aiming to release it in early June. So now I’m going to go back to Michael and casually mention the Ogilvy taxonomy like I’ve always known what it is.

Jen Leonard: You’re welcome, Michael.

What Just Happened: OpenAI's Gets a Second CEO and Google’s Wake-Up Call

Bridget McCormack: All right, moving to What Just Happened. I think we're getting to the point where so much is happening that we just try to pick a couple of the most important things. Otherwise, it would take over the whole podcast because things are moving so fast. So what just happened?

Jen Leonard: Yeah, so this week, OpenAI's CEO, the famous Sam Altman, is still in charge. But this week, they announced they’re bringing in Fidji Simo, the CEO of Instacart. She’s going to take on a major leadership role at OpenAI. I believe her title is CEO of Applications.

Bridget McCormack: Which is weird—because there usually aren't two CEOs. So what it sounds like is that they're dividing responsibilities. She'll be responsible for product and applications—making sure people like and use the stuff OpenAI is building. Sam will focus on infrastructure and research—pushing toward AGI.

Jen Leonard: Makes sense, and she has that background—she helped modernize the Facebook app when she was there. She’s also been on the Shopify board, and Shopify did some experimentation with ChatGPT integrations. So this could be OpenAI moving into a much more commercial phase.

Bridget McCormack: Yep. The big question is—are we going to get ads?

Jen Leonard: I hope not. But Sam Altman has also been working with a former Apple hardware lead—the person behind the simplicity of Apple’s product design. So hopefully, they’ll avoid the kind of ad disruption that defines a lot of the current internet experience.

Bridget McCormack: She was already on their board, too, since the reorganized board came together last year. But her CEO title makes me think Sam might be setting things up to do the next thing. And I’ve been seeing a lot about his “orb” eyeball company lately.

Jen Leonard: I don’t think that’s what he officially calls it, but that's the one. It's part of this project to prepare for a world transformed by AGI. The idea is that we’ll need a way to verify people online, and also potentially distribute universal basic income. His team’s been traveling around the world scanning people's retinas with this orb that becomes their digital identity.

Bridget McCormack: Like your fingerprint—but more unique. I think it’ll also become a payment method, ID card, passport... Honestly, Delta and Clear have already scanned my orb. I'm basically ready to let anyone scan my eye if it gets me through security faster.

Jen Leonard: And we’ll probably have this phase where everyone resists, and then suddenly it’s like—oh well, this is convenient. But this connects back to OpenAI’s restructuring. They had been trying to shift to a purely for-profit company, got some legal pushback, and gave up on that. So now they’re moving toward becoming a public benefit corporation instead. And reading between the lines—it makes you wonder whether Sam's not that interested in the corporate structure anymore. Maybe he's handing off the day-to-day so he can focus on other things.

Bridget McCormack: Could be. It's clearly an intense job. And all of it together does suggest they're preparing for the next phase. Which makes me think: maybe they feel confident AGI is close. Not that it's all tied down yet—but they're signaling, “We’re not that far off.”

Jen Leonard: And while all of that’s happening, we also got interesting information out of the DOJ’s antitrust trial against Google.

Bridget McCormack: Yes! So this is the second major antitrust case DOJ has brought against Google. This one’s focused on their advertising business. A key witness testified—Apple’s SVP of Services, I think his name is Eddy Cue. He said that, for the first time in two decades, Google search usage on Apple mobile devices has dropped. He attributed that drop to people using ChatGPT and Perplexity. Because these reasoning models just give you the answer. You don’t have to scroll through ten blue links.

You ask O3 a question, and it goes and reads the blue links for you—synthesizes the info and gives you the answer you were going to piece together anyway. So it’s no surprise that user behavior is shifting. Google’s going to have to figure out how to adapt. They’re probably going to have to double down on Gemini, their AI model.

And right after that courtroom testimony, Google’s stock dropped 7%. It rebounded a bit—but the market clearly sees that shift coming. And I still think Google has huge strengths—Waymo, YouTube, enterprise data.

Jen Leonard: But their core search experience is definitely under threat.

Bridget McCormack: It’s funny—I like Gemini. But I still end up using ChatGPT most of the time. I don’t even know why. Gemini’s integration with Gmail, Google Docs, and Calendar should be a huge advantage. But OpenAI still feels ahead in user experience.

Jen Leonard: I always have a ChatGPT tab open. I forget to check Gemini, and even though Claude used to be my favorite writing tool. I haven’t used it in weeks. What was the last thing you searched on Google?

Bridget McCormack: I was trying to figure out if a restaurant in Ann Arbor was open on Sunday night after a day of travel. But if it’s open-ended? I go straight to ChatGPT.

Jen Leonard: Me too. If I want a Wikipedia page or one specific link, I’ll use Google. But it’s not surprising to me that their search history has dropped and will continue to drop as more people use ChatGPT.

Garfield AI: The U.K. Just Approved Garfield Law—A Fully AI-Driven Law Firm

Jen Leonard: So a lot of interesting stuff is happening in the land of AI. And there’s also some really interesting stuff happening in our own profession, which is slower to change than the tech landscape, but not immune to change. And we are hearing news about one of the first AI-native law firms—not in the U.S., but in the U.K. So, Bridget, can you tell us what’s happening across the pond?

Bridget McCormack: Yeah, so this was an interesting story that the AAA’s general counsel, Sasha Carbone—who’s amazing—sent to me over the weekend. This law firm in England and Wales called Garfield Law, also known as Garfield AI, was authorized by the regulator in England and Wales, which is apparently called the Solicitors Regulation Authority—the SRA—to provide regulated legal services exclusively through AI.

So, not “lawyers using AI,” but AI directly providing legal services. It’s a small use case—it’s designed for debt recovery and small claims court. It hasn’t been approved for legal research or suggesting case law. I think these are cases largely decided on facts. But still—it’s a fully AI lawyer providing legal services. It feels like a small use case, but on the other hand, it’s a huge deal, because we’re kind of far from that in the U.S.

Jen Leonard: Yeah, and one of the things I thought was really interesting—it’s a collaboration between a lawyer and a hard scientist, someone with a deep background in tech. And I think the law firm was born out of frustration that one of their relatives had as a small business, chasing after a debt someone owed them.

I love ideas that are born of someone’s real pain, because it becomes a human-centric solution. It’s not “let’s build a business to make money.” It’s, “how do we fix this thing that’s broken?”

The regulators spent eight months reviewing the plan and coming up with guardrails. Which to me, sounds like a minute compared to how long we’ve spent in the U.S. debating, sandboxing and thinking about structures. And the article painted the eight months like it was a really long time, but by American standards? That’s lightning fast.

And one of the guardrails I loved was: it can’t generate case law. So it’s not hallucinating fake citations. I love that. Because the number one thing I hear from lawyers—especially litigators—is, “I won’t use AI because it makes up cases.”

So this eliminates that barrier. And it’s in a context where it doesn’t need case law anyway—just get the facts, move things forward. Let’s put that issue to the side and watch how this evolves.

Bridget McCormack: Yeah. And then it’s not that far a leap to the next use case. There are tons of high-volume dockets in state courts that deal with one problem over and over—evictions, traffic court, debt collection—where even if there is legal doctrine in the background, it’s not hard to train a system on the rules.

This feels like a great first step to show how AI legal services can impact real people, in real volume—and with outcomes we can measure.

Jen Leonard: Yeah, and I bet the founders could’ve built a GPT in ten minutes that does this and just skipped all the oversight. But they wanted to do it above-board.

Which raises a bigger point: There are so many different use cases, and if you find an area where the law is straightforward, and you could train an AI model to generate the documentation, how do regulatory structures keep pace to regulate AI models in the U.S.

Bridget McCormack: Richard Susskind says in his new book about AI that we need the regulatory bodies to keep pace with AI. He's like, we need cycles of regulation that are weeks, not years. And I get why he thinks that. I think he's right, but it seems so impossible, right? I don't know any regulatory body in the U.S. that's moving that quickly.

Jen Leonard: No. And I've thought this for a long time about AI generally. I mean, obviously, we're now in a deregulatory environment. So there's not effort focused on this. But even if there were—if your lawmakers understood it, if everybody agreed on what the sort of social norms would be that govern it—I still don't know how you could conceivably keep pace. It’s such a shape-shifter. It changes so quickly.

And then you think—in the U.S., we have all these different state-by-state jurisdictions. You and I have met with many of these bodies. I would say that it's not capable of moving that quickly, just by design. But to your point, I'm also not sure what fast enough would look like.

Bridget McCormack: Yeah. And I don’t think the people who make up those regulatory bodies are the problem. They've got a lot of other responsibilities, too. So it’s not a critique of individuals. But the regulatory construct itself just isn’t fit for this purpose.

Lots of states have been trying to figure out how to think about paraprofessionals for a long time. And a number of states have opened either a regulatory sandbox or even a court-approved pilot to allow paraprofessionals to provide legal services in really specific use cases—some in family law, some in eviction cases, some in debt collection. But that’s only happened in a handful of states.. And that’s despite the fact that everybody kind of agrees this is needed.

And in the meantime, this technology has kind of leapfrogged the whole paraprofessional idea. The solution that paraprofessionals represent—a more affordable option for basic legal help—has already been outpaced by AI. All of a sudden, we’ve got this more immediate, more scalable solution that lives on everyone’s hand computer. It’s available now, and that changes the conversation.

Jen Leonard: If you take the U.S. state-by-state regulatory framework at face value, and had a law student build a personalized GPT for each state in a given area of law—train it on local statutes and common filings, you’ve got a functioning legal assistant at your fingertips.

Jen Leonard: But I think Michael Sander has said this before—that if you sat there and used Control-F, you could have flown through all of that work. The problem was, you'd go to the partners and explain Control-F, and they’d say, “Go back and look word by word,” because that’s their business model.

And this is like Control-F—obviously—on a huge scale. And students are going to be using this from the time they’re in high school now. And I just don’t know how you stem that tide—having them come into a profession where they’re now told, “Just look away. This doesn’t exist,” when their whole lives are infused with it. It just seems so bizarre to me.

Bridget McCormack: I mean, it’s part of how it all just becomes like background noise. Recently, I was on a panel on AI and one of my co-panelists was saying, “We didn’t start by defining AI. And all AI is, is software.” Like, eventually it will just blend into our background and we’ll be used to it. Like, think of all the software in your life that you never think about—how it works or stress about it. Eventually, that will be true here as well.

And part of how it gets to be that way is these young people now coming out and entering the job market—the legal job market, the other job markets—are going to be used to using it for everything.

Jen Leonard: But to your point, I just don’t know what a profession would look like if you told a young person they can’t use artificial intelligence in three years from now. So this is definitely the way of the future. And again, the UK has led the way in innovating in legal services before.

And we could talk about alternative business structures, which they’ve had there for 12 years, that allow lawyers to own legal services organizations with people who are not lawyers—which we only have here now in one jurisdiction. And it was just approved in 2024.

Bridget McCormack: Yes, Arizona is, I think, the only state so far to allow for approval of alternative business structures. And they approved KPMG to launch KPMG Law, which is the first of the Big Four accounting firms to start offering legal services.

And when they approved it, I thought, “Okay, here we go—the dam opens and we’ll see all these new applications and efficient products and services.” But I haven’t heard a lot about it yet. It's hard when you can only practice in Arizona, I think, to make big strides. But I assume there’ll be others. And I assume this technology will push others to move more quickly. Like, KPMG will get some other Arizona competitors, right? And we’ll sort of see what happens with that. And then there’ll be the version of Uber out there, right? There’ll be some people building products that don’t even understand that they’re not supposed to do that, right?

Jen Leonard: And then it becomes a ridiculous game of whack-a-mole. And then the question is: What is your role in regulating?

And one of the things I thought was interesting in the article about Garfield Law—was just this idea that they structured regulation—or they think of regulation—in the U.K. as both protecting the public from risk and leveraging technology to expand access to justice. And I think in the U.S., we’re so focused on the first, but not balanced as much with the second.

Bridget McCormack: More evidence of that is in the different ways the two countries—I mean, in the U.S., of course, we have 50 models of it—but still, they’re more consistent than not across the states in how we define the unauthorized practice of law.

In the UK, it’s really limited to what happens in a courthouse, in a courtroom. But if you want to help somebody with a legal problem in your kitchen, and they’re your neighbor… obviously that happens in the United States. And now happens more than ever, with everybody’s access to much more information and tools to find it. But technically, you're in violation of many state criminal statutes when you do that.

So that reflects a different way of coming at the problem.

Bionic Boutiques and the Future of Legal Business Models

Jen Leonard: The other thing that the story made me think of was our mutual friend, Jae Um who’s done a lot of great reporting in the last couple of months for The American Lawyer. She puts out the power rankings for AmLaw every year. And she wrote about Bionic Boutiques, this is her prediction that in the next five years, you will see Big Law firm partners who just want to practice at the top of their license and do really sophisticated work.

They no longer really need the brand name of a big firm. They’ll start their own firm. And she calls them Bionic Boutiques, with a handful of other professionals they want to work with, lawyers they want to work with. And smartly, Jae is always thinking about: How do we work within the bounds of the system that we have to make change?

And here, in light of the regulation, the way you do that is, you automate everything on the business side and allow the lawyers to just practice law. And I know you were really taken as well with some of Jae’s predictions for the future.

Bridget McCormack: I think it’s a great piece. I recommend reading it. She thinks that we’re underestimating how much change we’ll see in the medium to long term and maybe overestimating how much change we’ll see in the short term. But the change that she predicts—these billion-dollar law firms..

Jen Leonard: With $10,000-an-hour billable rates and $10 million profit per partner

Bridget McCormack: But with the army of AI agents to do kind of the lower-level work. If you’re the Bionic Boutique partner, and you have a bunch of agents who you can train on all of your knowledge that’s on all of your prior work product, then you can have a subscription model for all of your clients to get information up to a certain point until they need to talk to you, the bionic lawyer.

There are new ways of leveraging, in other words. The old leverage model does feel really old now, given this technology. But there are new leverage models. And I don’t know if that’s good or bad.

Jen Leonard: I guess it depends on who you are, right? If you’re the $10 million profit lawyer, it’s good. If you’re a junior lawyer… I don’t know.

I just interviewed Jae, and she talked about how she makes her predictions in a way that changes the framing of the conversation to be a little audacious, so people start talking about things in new ways. But I don’t actually think it’s that far off.

If you’re the general counsel of a major global company, and you have a major tax issue, you’d probably be happy to pay whatever it costs to get the best tax attorney in the world to advise you. What you don’t want to spend money on is the entire leverage model that’s baked into that partner’s current Big Law structure.

And if you’re the Big Law partner? You’re probably feeling the strain of that model too. You’d love to break free and just practice.

Bridget McCormack: And I assume if you’re the billion-dollar bionic lawyer, you have all these AI agents to help test your ideas. If you’re trying to solve that extremely high-value tax problem, you have AI agents helping you stress-test the solution.

Jen Leonard: And that’s probably the regulatory-safe business bet that a young junior person could help with, right? Come in and be the head of AI transformation at your bionic boutique-to-be. And work with the smartest tax attorney in the world to figure out what agents they need—and then go build them.

I wonder if it connects back to anything we talked about earlier with OpenAI, the acceleration of commercialization, people like Sam Altman possibly moving on, and what’s happening in the legal landscape. Do you think there’s a connection between those stories?

Bridget McCormack: It feels to me like the technologists know what’s coming. We’ve been talking about this since those conversations in December, January, February—about AGI maybe being closer than we thought. And now it’s almost like everyone’s preparing for the next phase of this, even though not everyone’s even caught up to the current phase.

I still think law is going to be a little slower to fully have it all move into the background—so it’s just software that we don’t think about. But not forever behind.

I do see OpenAI is now doing these benchmarks across different disciplines. Ethan Mollick posted about one in the latest healthcare benchmark. And we talked a number of episodes ago about that one study that evaluated doctors alone, doctors using ChatGPT, and just ChatGPT.

And the latest reasoning models are just far and away better than either the doctors using it or doctors alone. Like—not even close.

To me, that just means medical schools probably need to figure out how to teach future professionals to use the technology to provide better service. But the fact that OpenAI is benchmarking across industries—that feels like another sign. Like, “Hey folks. This is happening. It’s time to come along for the ride.”

Jen Leonard: I think you’re right. Legal is always going to be slower to change, and when we talk about legal being slow to change, what we’ve mostly been talking about is the people who currently populate legal.

And we’re slower to change for lots of reasons—good, bad, neutral. And so I wouldn’t be surprised if people who are not part of our profession, people like the new CEO of Applications at OpenAI are thinking, “How do we get into the legal market?” And maybe, because we’re so slow to change, that makes us even more vulnerable to disruption there. We might not even see it happen or see it coming.

Jen Leonard: Is this going to be another Uber moment? Where the disruption just happens, and by the time we try to react, it’s already too late?

And then you’ve got this new sharp talent at the head of OpenAI, and they’re moving into a totally commercial space, and they’re a public benefit corporation. I can’t imagine anything that would benefit the public more than actually having access to legal services.

All of those shifts, to me, mean we could be watching the profession change from the sidelines because we weren’t pressured enough, or open enough, to innovation all along.

Bridget McCormack: It is inevitable. I don’t hear people fight anymore about whether it’s real or happening. But if it’s inevitable, the only real question is: Who do you want making decisions about the future of our profession? Shouldn’t it be us? People who care about the profession and the role it plays in a constitutional democracy.