AI’s Impact on the Attorney-Client Relationship

Your clients are already using AI. The question is, are they using it wisely?

In this episode of Mintz On Air: Practical Policies, host Jen Rubin is joined by Steve Osborn, Managing Member of Mintz’s San Francisco office, for a conversation exploring:

• How clients are using AI to evaluate legal advice and challenge traditional notions of attorney trust
• The risks of sharing sensitive business or legal information with AI tools
• Practical ways attorneys can use AI to deliver more value to clients
• Why trust, professional judgment, and accountability continue to distinguish legal counsel from AI-generated advice

This episode offers practical insights on using AI responsibly, collaborating effectively with clients, and preserving trust while navigating technological change.

Mintz On Air: Practical Policies — AI’s Impact on the Attorney-Client Relationship

Jen Rubin (JR): Welcome to Mintz On Air: Practical Policies. Today’s topic: AI’s Impact on the Attorney-Client Relationship. I’m Jen Rubin, a Member of the Mintz Employment Group with the San Diego–based Bicoastal Employment Practice representing management, executives, and corporate boards, and I enjoy providing creative solutions to workplace problems. Thank you for joining Mintz On Air. 

Today I’m joined by Steve Osborn, Managing Member of the firm’s San Francisco office. Steve is a versatile corporate attorney who draws on his in-house legal and business management experience to advise clients on transactional, corporate finance, and general corporate matters. Steve focuses on middle-market mergers and acquisitions and regularly advises clients on exit strategies. He represents public, private, and emerging-growth companies in a variety of industries, including life sciences, technology, digital health, and retail and consumer products. I know Steve is keenly interested in how AI is changing our profession, and I thought our listeners might benefit from some of those insights. Thanks for joining us today, Steve.

Steve Osborn (SO): Thank you, Jen. I appreciate you inviting me to the podcast. I’ve been enjoying listening; now I’m a guest and I appreciate that. Thank you also for that generous, thorough introduction that made me sound very impressive.

JR: Let’s start by my quoting something you said recently that left an impression on me. You said that one of the significant ways AI will impact the delivery of legal services is that lawyers will no longer be viewed as professionals who use technology to deliver legal services. Instead, AI will make law firms into tech companies that deliver legal services. Now, I know you did not mean to dehumanize the human aspect of our profession. But it’s undeniably true that AI is changing how we deliver legal services. I’m looking forward to this conversation because I want to talk about a subset of that paradigm, which is how AI is impacting and changing the relationship between lawyers and clients. What does this mean for the practice of law? How is AI impacting how we relate to our clients and provide professional services to them?

How Is AI Changing the Attorney-Client Relationship?

SO: It’s true that I believe law firms are going to have to adapt to this new technology. Over some period of time, I do believe law firms will look more like tech firms providing legal services. That’ll be tough because in the legal industry, to put it in AI terms, adaptation is not a built-in feature. It’s more like a bug fix we’ve been avoiding for centuries. If you think about it, the way we practice law today is not too different from how Alexander Hamilton did it 300 years ago. It’s an apprentice model, we bill by the hour. That’ll need to change and we’ll need to adapt. But interestingly, clients will also need to adapt, and the working relationship will have to adapt. So there’s going to be a lot of change in the legal industry. We’ll have to adjust to that and it will not come easy. 

When it comes to using AI, one of the things about this product that’s amazing is it’s a consumer product. Our clients are using it right now. Anyone can get on the chatbot and ask a legal question and get some response out of it. And we’re seeing more and more that clients are doing a couple of things with AI. One, they’re trying to educate themselves before they get on the phone to have a conversation with us. Two, they’re checking the work we’re giving them against AI and then using that to challenge us. An interesting thing about that is that the idea of trust, which is one of the core components of working with a lawyer, is coming into question by the AI because the AI is providing responses that maybe question the legal work we gave. I find clients sometimes trust that more than they trust my advice, or that’s an inclination they have. And it does cause a bit of stress between lawyers and their clients. How do you feel about that, Jen?

JR: I think the term you use when we’ve talked about this is the “gamification” of the attorney-client relationship. Clients are challenging us with information they’re finding — which I happen to think is great, to have an educated client. I will never object to a client thinking about the problem and thinking it through before they come to me. But sometimes I feel like I’m a spell checker, right? Like I’m getting checked by the AI. I want to know that the value the client is getting is from the actual relationship. So in a way it’s a gamification. And I have to be truthful because I think it’s kind of fun. I really do. There’s an enjoyment, it’s adding this whole aspect to our practice that I find very enjoyable.

JR: Imagine, Steve, you meet a pair of lawyers — their names are Larry and Lisa Luddite. These are lawyers who don’t embrace AI. They don’t learn how to use it. They don’t accept it. They’re very skeptical. As you know, Steve, lawyers are very change resistant. We’re trained to be thoughtful and logical and to look around corners. 

SO: And trained to be skeptical, by the way.

JR: Exactly. A little bit of skepticism is great, right? But what advice can we give to lawyers about how to set that skepticism aside and embrace this technology and make it a beneficial part of the attorney-client relationship?

Why Should Lawyers Embrace AI Instead of Resisting It?

SO: One of the things about lawyers is that, to be a great lawyer, you have to be curious and you have to be willing to engage in critical thinking. The AI tool, when used appropriately, can be a great aid in helping you understand things you’re curious about that you may not know about before — maybe something around the edge of the advice you’re offering. I might be able to say to AI, “Can you explain to me why the noncompete provisions are drafted this way in these states?” Something like that, where I can get a little more educated myself to help advise a client at a higher level. That a great aspect. As you say, it’s exciting as a lawyer to have a tool that allows you to explore that curiosity. It’s like having a couple of your partners in your pocket.

The AI is an aid, it’s not the work itself.

One of the risks of putting in a search term into AI is that the AI is going to come back and give you what you want to hear. I have great stories and there’s been great stories in the press about how if you don’t use AI appropriately, you might end up going down the wrong path, because it’s going to follow what you’re saying and try to give you the answer you want to hear. So you have to use your critical thinking skills when using the tool. But if I have to answer a client quickly on something, I can engage with an AI tool and get that kind of critical thinking. Then I could come back to my client in a short amount of time with not only the good ideas I came up with, but some other ideas that maybe AI was able to generate. I still think you need to translate that work. The AI is an aid, it’s not the work itself. You translate that, but maybe one of the ideas that AI was able to come up with in this critical thinking engagement is really good. I add that to the mix of things I’m presenting to the client. For lawyers, this is a wonderful tool to have. 

How Should Lawyers Respond to AI-Generated Work Product From Clients?

SO: A lot of partners come to me and say, “I need to get trained on the AI before I use it.” The interesting thing is, that was true about LexisNexis and Westlaw and some of the other technical products in our space, but it’s less true about where we are in AI right now. A lot of where we are in AI is, it’s just an assistant. You can go to the machine and type something in and start engaging with it. Of course we are building more sophisticated tools and workspaces and repeatable models and things that are more advanced in the AI world to deliver client service, but at the basic level, it is a consumer product and it is easy to start training yourself by using it. I always tell our partners, “Let’s go to your desk and sit down and type in a question. And now you’re trained on AI. Go have fun and do the things we just talked about.” What are your thoughts, Jen?

Not only do lawyers need to learn how to use AI, but they need to learn how to talk to their clients about the work product the client is sending.

JR: There’s a fundamental difference when you and I have a conversation and we kick things around and we have a collegiality, versus when a client sends me something they probably used AI to create. I may look at it and I may say, “I’m not so sure about this.” I know the term is AI slop. I hate that term, it’s kind of a nasty term. But anyway, maybe you’ll get work product that’s not the work product you would produce as a lawyer, but the client is presenting it. It puts you in an interesting situation because maybe you’re on a back foot with the client when you receive that. You have to learn as a lawyer how to talk to your client about that, in a diplomatic way, of course, always courteous and professional. Not only do lawyers need to learn how to use AI, but they need to learn how to talk to their clients about the work product the client is sending. That’s an interesting case that people aren’t thinking about: How they’re talking to their clients about it. I’m interested in any suggestions you have, Steve, in terms of what AI is good for as a starting point or an ending point, and how you talk to clients about how they’re using it.

SO: I’ve had two clients this week try to help us — in their minds help us — by pulling together a lot of information. These are both in the commercial contracting area. The clients want to rebuild their commercial contract portfolio so that it applies to their business as it is today and it can be flexible and used in a variety of ways. That’s two clients who have come to me this week. They both are very tech savvy. So they prepared long briefs for us and pulled together a lot of information. I thought that was helpful, but they thought it would somehow turn a fairly sophisticated project into something that was easy for us to do. Their expectation was that they put all this work in and used AI to pull all this together and give us a head start. And I had to say, “I hate to tell you this. You saved us one phone call because I would have gotten on the phone and asked you for all this stuff. But the hard work here still needs to be done. We still need to do the critical thinking. We still need to understand your business. We still need to make sure we understand the nuances of how your business applies to privacy laws in this country and other countries. There’s still a lot of work that needs to go on.” The AI, in that context, is not actually saving the kind of time that clients think. And that has been surprising to those clients. 

Also sometimes when a client uses AI to prepare something, the challenge is that it actually takes more time for us to unwind what they’ve done. The tools right now used by clients as a proxy for skinnying down the legal work are not working effectively. I’d prefer the client to call me up and say, “How can we work together to use AI to make this project not a $20,000 project, but a $10,000 project?” I’d like them to talk to me about that before they run off and put a lot of stuff together and convince themselves they’ve got a $10,000 project and me having to give the bad news after the fact.

JR: That’s a great segue into this point: What are clients doing with AI and talking to AI about and prompting AI to do that creates issues? You know what I’m talking about, Steve.

Does Using AI Put Attorney-Client Privilege at Risk?

SO: Yeah. Let’s put it into a TV crime story framing. If you committed fraud or thought you committed fraud or think you might be accused of committing fraud, and you go to the AI chatbot and you say, “What is fraud?” Then it tells you what fraud is and you’re like, “Huh, maybe I did.” What you’ve done is you’ve created a great record for the prosecutors because none of that is attorney-client privileged. People don’t realize that when they’re asking AI for legal advice — which a lot of our clients are doing — all of that is discoverable. People usually know their Google searches are discoverable, right? Similarly, we’ve learned to be careful with email. But I don’t think clients have learned to be careful with their AI searches. How are you advising clients, Jen? Because it comes up a lot in your area, where CEOs are riffing on employment issues before they even come to you, trying to make decisions.

People don’t realize that when they’re asking AI for legal advice — which a lot of our clients are doing — all of that is discoverable.

JR: Exactly. The beauty of AI is the ease, right? As you said, it’s a consumer product. Anybody can sit down and put inquiries in. But what we’re saying to our clients is, “Come to us and work with us. There’s a way to do this efficiently. Because when you come and ask us these questions, it’s privileged.” And the attorney-client privilege is a very useful privilege. It goes without saying, you want people to be able to be comfortable seeking legal advice. That’s the whole paradigm of the law, to encourage people to go to a lawyer and say, “I’d like to know what my legal rights are, I’d like to know the legal way of doing this, I’d like to understand how we approach this issue.” And AI can help us reach those conclusions and answer those questions. When those questions are asked in tandem with your counsel, it’s going to be privileged — but if you are doing these inquiries on your own, it’s not privileged. 

Another thing to remember, if you’re using an open-loop system or a system that’s going out and training other AI models, you’re putting information out there into the AI world that other people can access. And that creates confidentiality concerns. 

JR: Let me ask you this, Steve, because this is something else you mentioned previously: All of us know AI’s sycophantic way of saying you’re wonderful, you’re the best, you’re so smart, I’m so glad you asked me that question. How do you handle that when you’re talking to clients about these things? Because it’s becoming a loop of wonderfulness.

Can AI Replace a Lawyer's Trust, Judgment, and Accountability?

SO: When you look at adoption of AI tools in professions, it’s clear that coding is a place where AI has been adopted and is making a big impact. I think the second or third most prevalent area is actually legal. AI is already making a big impact on how we work, the amount of work we get, whether clients come to us or whether they do it on their own. So it’s good for us to think about first principles here. What is the reason you would come to a lawyer like us? I had a nice conversation with a client the other day, he was very savvy on this stuff, and he said, “Well, Steve, I’m going to come to you no matter what the AI says. I’m going to ask your advice because I trust you. I know you have my back because we’ve been through it before and you’ve proven to me that you deserve my trust.” They come to us for our judgment because we’ve seen things like this before. We understand the law and we understand the business rationale of what they’re trying to do. They come to us for the critical thinking. It’s very hard to get that right now from AI. Clients don’t bring us their easiest problems, right Jen? They bring us their hardest problems. And you and I love that; that’s why we still are in private practice and do what we do. I want the hardest problem you have and I want to work with you on thinking it through. That nuance, that level of critical thinking needed to solve hard problems — AI cannot do that.

Lawyers stand behind the decisions we make. That kind of professional insurance is something AI can’t replicate.

And then of course there’s an accountability piece. We’re a human layer of accountability and assurance. Lawyers stand behind the decisions we make. That kind of professional insurance is something AI can’t replicate. Clients use lawyers for trust, judgment, critical thinking, and accountability. Some of those things might get replaced by AI over time, but we’re not in any danger of that happening in the near future. So there’s a lot of value in using lawyers, and AI is going to be a tool that helps us improve our legal work, improve the speed, and lower costs over time. 

The thing we talked about earlier that I want to come back to in this area is, we shouldn’t be sneaking around. The client shouldn’t be sneaking around using AI and not telling us. We shouldn’t be sneaking around and using AI and not telling the clients. We don’t do that by the way, but we shouldn’t be doing that. I tell the Associates, don’t be sneaking around using the AI. Tell me when you use AI. Tell me what prompts you did. Teach me how to use it better. Let’s talk honestly about how we’re using it. And I think that’s true with clients. We should be coming to clients saying, “How can we use AI tools to improve this process?”

I had a situation recently where we needed to respond to 24 different counterparties — not all of whom, by the way, were lawyered up. Some did have lawyers, and some used their own versions of ChatGPT to respond to us. I was getting long emails from them where we needed to respond to that. So I worked with the client to say, “How do we respond consistently but also compassionately and helpfully to each of these clients to get them to a place where we can resolve the issue we have with them.” And I said, “Let’s use AI tools to create a database of the questions that are asked and the answers we’re going to give, and then proof those answers and then get AI to run those across all of these letters. And then we’ll proof those letters and send them out.” In the old days, that was probably three people and four or five days of work. And we were able to do that in one day with one person, with one senior lawyer. So there’s an example where the client and I found a way to save a tremendous amount of money for the client. But we worked on it together. We did not work on this separately. My big encouragement to clients is: don’t try to use tools and then connect to your lawyer. Connect to your lawyer first and talk about how you can use AI tools together to save money, time, and make the product better. Of course, if your lawyer doesn’t know what you’re talking about, then you’ve got the wrong lawyer in this modern era.

JR: To riff on that and to further tease out this concept of accountability, we train for a long time to do what we do. We’re holding ourselves out as professionals. We’re constantly learning and providing additional education. We’re looking at what’s going on in the marketplace. All of these things influence our judgment. You talk about trust, right? Clients come to us, they trust us, because we’re not just going to turn to them and say, you’re wonderful, you’re the best, there’s nothing wrong with what you’re asking me. Hopefully we can talk to our clients and redirect them when we think they’re headed in the wrong direction. That’s the accountability piece. We are standing behind the advice we give. It relates to this concept of vouching. When the client goes back, the board and the C-suite are going to want to know, or maybe it’s opposing counsel or a judge who’s going to want to know, “How did you come up with that answer, and do you stand behind that work?” Just saying “I asked AI” is not enough. This all goes to accountability and us standing behind the work we do — the work we do in collaboration with the clients and with these AI tools. That’s where you and I are coming out here, that all of these things need to work together and we all need to learn, because it’s a whole new world every day. And this is what we need to be moving our practice toward. What makes our relationships with our clients so strong is the fact that we’re willing to do these things or learn these things. 

JR: So this is the Practical Policies podcast. What advice, Steve, can you give to lawyers, in-house counsel out there, or even clients who are tuning in? What are some things we should be doing now as we’re going through this rapid change in our practice?

What Should Lawyers and Clients Do to Use AI Responsibly?

SO: One thing is that you need to be engaging with these tools at some level and you need to be doing that in a responsible way. For our clients, they should be employing one of these systems and they should do it in a professional way that allows for the information to be confidential and not to be used for other training purposes. Clients who are not lawyers need to realize that this is not attorney-client privileged, even if you’re asking legal type questions. If you find yourself asking the AI legal questions and you’re not a lawyer, you should stop right now. I don’t think that’s obvious to people. It’s new technology, it’s exciting, it’s fun, and we’re rushing ahead sometimes before we think about it. We should be working together with clients on this because the tools are powerful, they’re getting better, they’ll never be worse than they are today. Things we couldn’t do with AI last year, we can do this year, and things we can’t do right now, we probably will be able to do in a year from now. So we need to be working with our clients to iterate on those things and realize that the outside law firm, lawyers in-house, and business people play an important role in this process of providing legal advice. If you’re going to solve tough problems, you’re going to need the machine and the human. You’re going to need somebody who’s well trained in these areas. Law firms like us are thinking about that. We’re thinking about how to train our own people and how to use these tools better, but we’re also thinking about how to train our younger attorneys, and we’re adapting. Firms like us have the ability to adapt quickly. Firms like Mintz have always collected curious, growth-minded people. My colleagues are constantly pushing the limits and bringing new ideas to me, and my clients are too.

If you find yourself asking the AI legal questions and you’re not a lawyer, you should stop right now.

So I think that’s the right advice: Have your colleagues and the people around you push you to use these tools and learn how to use them and then connect with the professionals in your life to use them effectively and not go it alone. That’s my advice.

JR: That’s great advice. I love the focus on curiosity — that’s what led to my doing these podcasts, because I am curious and I know you are as well, Stephen. You’re doing a lot of this work and I enjoy watching the posts you’ve been putting out on LinkedIn and elsewhere. Being curious is what keeps us fresh. It’s what keeps us human, to be honest. It makes us wonderful professionals and partners for our clients, or at least I hope it does. So thank you, Steve, for joining me today. I really appreciate it. This has been a great discussion.

SO: Thank you for having me, Jen, and I agree. I always enjoy chatting with you.

JR: For more information about Mintz and our practice groups, as well as our thought leadership, you can visit the Insights Center at Mintz.com, or you can find our Mintz On Air: Practical Policies podcast on Spotify. I’m Jen Rubin. Thanks again for joining us.

Subscribe To Viewpoints