AI at the crossroads: The legal questions every business is facing

Transcript:

Gerard: Hi, this is Gerard Donovan at Reed Smith, and welcome to Tech Law Talks. AI raises legal questions that don't respect practice areas or geographic boundaries. So we built a global AI practice that doesn't either, with lawyers across practice groups and geographies. Let's get into it. First, to do quick intros, I'm Gerard Donovan, a partner in Reed Smith's IP group in Washington, DC, and a co-head of our global AI practice. I'm a computer engineer turned IP litigator and counselor. Which means I spend most of my time at the intersection of emerging technologies like AI and IP disputes like patents, trade secrets, copyrights, and and all of the issues those raise. Tyler, you wanna go next?

Tyler: Sure, yeah, thanks Gerard. My name's Tyler Thompson. I'm a partner in the Denver office. I'm one of the co-chairs of Reed Smith's Global AI practice. I originally was a data privacy, cybersecurity, and tech transaction attorney. And while I still do a lot of that, now the practice has evolved to be heavy in AI as well. And Nick, I'll hand it over to you.

Nick: Thank you. I am a partner in London within the Entertainment and Media Industry Group, focusing predominantly on music and video games matters, including advising on how emerging technologies like AI are affecting those industries.

Gerard: And here's the idea behind the show. AI is no longer experimental. It's embedded in how business operate, how content gets made, how products get to market. But legal questions it raises don't stay in a single lane. A single AI deployment can simultaneously be an IP problem, a regulatory problem, a commercial rights problem, or raise a dozen other areas of

Tyler: Yeah, that's right. I mean, too often we're discussing AI as if it's one narrow specialty area, but in reality, AI issues, AI risk really crosses a lot of different what we view as kind of traditional practice areas or traditional businesses, right? so reality the hardest questions are where some of those overlaps happen. So where privacy over intersects with IP, where product counseling intersects with things like data security in the AI space. Anything in that gray area is gonna be difficult. So and where you think about, hey, where we have this cool AI use case, where does that meet the liability for AI? Those are the interesting areas that we want to dive into.

Nick: That's right. So in each episode we'll we'll take an AI issue and we'll look at it from multiple angles at once. And then the three of us will host the episodes. We'll bring in colleagues from across the firm and occasionally special guests from industry to do a deep dive in their own corner of the AI world.

Gerard: Think of us as your standing panel. So today it's the the three of us setting the table, introducing the issues. and going forward we'll we'll bring in whoever can best unpack the issue of the day.

Nick: Yeah, so for this first conversation, we wanted to do something simple. Just introduce ourselves, stake out how each of us sees the AI landscape heading into the rest of the year, and surface the questions that we think matter most. So I guess consider this the map before the journey.

Tyler: Yeah, and we're certainly not going to agree on everything, but I think that's more of a feature than a bug. a lot of the regulation debate alone right now is very contentious where we think, you is regulation killing innovation or do we need to regulate? this is a question alone that a lot of reasonable people are fighting about.

Gerard: So let let's get into it. to open, a question to you know either of you, what's the biggest single shift you've seen in how clients are using AI over the past year and why does it matter?

Tyler: I can jump in. I mean, I think it's a it's a cultural shift really of you know when we first started counseling on really some of the new generative AI things, AI has been an issue for a long time, but well let's talk about with the rise of generative and LLM. a lot of my clients were very fearful of it and it was, hey, how do we put in a policy saying, no, we can't do this, don't use AI. No, now the the culture and maybe it's the demand from the top has changed so much it's basically, hey, we have to use this, we're hearing from our CEO from our board, our chief legal officer that we need to incorporate AI, how do we do that effectively? So I think that's the biggest change for me is just that cultural flip from being fearful of it to now, hey, we've gotta embrace it or we're gonna be left behind.

Nick: Yeah, I think I can echo that. the biggest single biggest sort of change I have seen is in some instances almost a 180 in how clients are approaching the use of AI from as early as the same time last year, where some clients were saying no AI at all, or extremely restricted use of AI, to now, only a few months later completely reversing that decision and going full in on on AI, requiring all employees and suppliers to demonstrate how they're using AI as part of the services. So you know that obviously brings with it some pretty significant requirements on internal governance and policies and procedures to be put in place to ensure that that can be done safely and responsibly. So that's probably the single biggest shift I've seen is is is that almost the attitude has changed.

Gerard: Yeah, I'll say the the biggest shift I've seen relates to those two. I mean, a lot of the clients I've worked with over the last fifteen years that were in the AI ML space were like real innovators, inventors, developers of stuff. And so ML was always like a niche thing. And now I'm working with, many clients that are inventing, creating things because as everyone's adopting AI, and especially as like capabilities and you know, a normal person can vibe code and create things. there's there's just so much more innovation that in your industry when you're using AI and your competitors can also use the same AI tools, there's a lot of motivation to not just use AI, but to develop systems to use it better than than everyone else. And so there's just a lot of innovation and exciting stuff. Let me ask a different question. What's the most common misconception you hear about AI and the law? What do you find yourself kind of hearing again and again from clients.

Nick: I mean for me it's it's any time when a a client sort of talks in absolute terms about anything. whether it's about training AI systems and whether that constitutes infringement or requires consent, whether AI generated works can be protected by copyright. Yeah, the truth is in each of those instances and probably most others, it's highly fact-dependent and there's no one size fits all answer and I do find that clients tend to talk about things in fairly broad strokes of well this this is fair use or the opposite this is this requires licensing uh whatever the case may be so I think that's probably the biggest misconception is it's almost in all instances never as clear cut as as being as black and white as that.

Tyler: Yeah, I'd say from both clients and the general public, I think there's a misconception. People say AI is not regulated, right? And I think maybe goes a little bit back to back to what Nick was just saying, people want to talk in absolutes. I get why folks say that, right? I mean, they're thought thinking of, well, is there kind of a broad, overarching federal piece of AI legislation? Uh no, but you know, plenty of existing areas, IP, you know, in privacy, for example, we've had privacy regulators come out and say, we don't think we state privacy regulators say we don't think we need a new AI law. the state because we think the privacy law that already exists in this state allows us plenty of enforcement capability and we're going to be exercising that in the AI space as well. Then you see a lot of states with niche kind of one-off AI laws in certain areas. So the ever I hear that kind of absolute like Nick said of well AI is not regulated, it's always not so fast. AI is actually in some use cases quite highly regulated. It's just kind of the overarching unified regulatory scheme that we don't have yet.

Gerard: I'll say in my space it's kind of similar to that, Tyler, that I I what I often hear is, conclusions about whether eh AI creations can or can't be protected, but from a a limited window of looking at, for example, copyright or looking at patent, but not not thinking about how the whole, mix of rights, name, image, likeness rights, trade secret rights, all this can impact what you can protect, what your risk is, and and how to make sense, how fact dependent it all is, because it's not when you're creating a lot of stuff, hey, this is copyrightable, this isn't. A lot of times, you know, there's aspects that have your creativity and those those can be protected. Other aspects can't. And you gotta dig into the facts and help them develop a program or understand the risk. really understanding what they're doing. It's not just we're using AI, so what's the answer? It's it's never that simple.

Tyler: Yeah, I think that's a really good transition into diving into IP. And I could start with a question that, you I think about a lot, but I'm wholly unqualified to answer. So maybe I'll kick it to you, Gerard. But really who owns an an AI generated work? And besides that $100 million question, what does the US, what does US law actually require right now for works like

Gerard: To give the the lawyer answer, it depends. And it depends on kind of what aspect of ownership you're looking at. The the question people are most often asking is copyright. and for that it it's gonna depend on if there is human creativity in the work. And so if you are designing something and your own, you creative expression is used to you know trigger what's being made. So say you're making a piece of art, putting it into an AI and then It's outputting something, based on that in additional prompts. Or you prompt and then, you take the output and you're modifying it. if if you're making code, you're rewriting aspects to incorporating into what you're doing. For those, there would be aspects that are are copyrightable, copyrighted by creating them because they are the human input. And there'd be aspects that are AI created, and so you can't protect that. So you gotta parse between the two. It gets a little more complicated or potentially gray when you're looking. I mean, the the copyright office's position is that if it's just prompting, then that's not enough. That's the creativity is from the AI. It's like hiring an artist to make something. The the artist is making the decisions, they're they're making it. so that's not copyrighted. It's kind of yet to be seen though, as AI moves forward and as the way you interact with systems change, whether there could be sufficient creativity to to give copyright even with prompting and not editing afterwards or providing other input. You also have copyright though, or sorry, you also have trade secret, you also have patents. I mean, for patents, if there's an invention, you if it's a human invention, you can still have patent rights for it. You can use AI as a tool to assist, just like you could have lap as a lab assistants and others, work with you. Uh and so as long as it's you coming up with it using AI that's patentable. If it's just a, hey, I have a problem, AI, what's the solution? And it spits out an answer, there's no patent rights there because you didn't come up with it. That that's AI came up with it. You might come up with an improvement that you can patent, but you can't patent what AI invents. Uh and then for trade secrets, this is kind of one of those, Tyler, your kind of point that you have areas of law that aren't AI specific but still apply to AI. Anything created by AI can be a trade secret if you're taking reasonable measures to keep it secret. The fact that it's created by AI doesn't change that. So you can have trade seeker rights and things that you're creating by a self-hosted model, by something on an enterprise where you have confidentiality terms. At the same time, if you're using a public model or if you're putting information into Chat GPT, just public version, then you're gonna have no rights to it at from a trade seeker perspective.

Nick: I guess sticking with patents and trade secrets. we're seeing a lot of disputes tied to AI. What kind of disputes or AI disputes do you expect will define the next couple of years? Je pointed to Gerard.

Gerard: Yeah, I'll I'll say I expect there's gonna be lot more and you know, I might be reading this wrong. I think copyright's gonna be the the one that has a lot of focus for the next two to three years. We're all waiting, I I think globally, we're paying attention to what's happening, you everywhere else, outside of the US too, to figure out the fair use, figure out, training data, figure out kind of as normal licensing schemes happen and just the value that's invested in AI. I think those are gonna continue to be litigated. There also right now is just in the US a huge surge of trade secret cases. And again, it's I think sort of tied to the value of this and the value of the the information that's being used and some of the people moving between companies. So those will be a a big surge, but I don't think those are gonna get quite the legal following, I guess, as the copyright cases.

Nick: Yeah, makes sense. And I remember you and I were on a on a panel at South by Southwest this last year. And I I remember you characterizing the role of a lawyer and how they should be playing their role in the client's strategy and sort of drawing this distinction between an architect and a gatekeeper. It'd be super interesting to understand what you meant by that.

Gerard: Yeah, I I think I mean we've seen clients more and more want to use AI. They want to figure out how they can, you make things more efficient, create new things. Uh and they they also need to mitigate risk. They also need to, you own as much as they can and and navigate all of this. And it's easy to say, hey, there's risk involved with all this. I mean, for example, copyright, this law is largely unsettled. We're starting to get some decisions and I think the more you understand how systems work, you can kind of apply existing law to those facts and have a good prediction, but we don't know. And if you're going to be, you very risk averse, you can sort of stick with the horse instead of moving to the car. And they want to understand how to how to move forward, how to take advantage of everything that AI offers. And so helping them find the right, risk mitigation approaches, the right ways to protect their IP so they can move forward understanding what they're doing, understanding the evolving law and and being strategic about it instead of being overly cautious. I mean caution's important, you want to understand the risks, but be able to calibrate and and move forward. Why don't we uh switch and Tyler, I have a question for you. you live in a world of regulatory uncertainty. If a client asks you today whether to build their compliance program around US rules or an EU framework, what would you tell them?

Tyler: Yeah, good question. I mean it's it's kind of neither and both at the same time, right? And I guess, you know, to get in this discussion, we're assuming that we're talking about a client that look some of the kind of more niche or US specific AI laws don't apply. Uh the EU AI AI Act doesn't directly apply, which besides like our developer clients, I think is still where the majority of clients are, but that's a shrinking base. But I think you have to look at both, right? So you're trying to figure out really what is what is reasonableness so that's what we're just worried about a lot right now what clients and companies are worried about is okay hey if my board if some future piece of litigation if the New York Times starts grilling me on what I'm doing with AI can we show that we put something reasonable in place and a great metric for that is looking at some of the rules in the United States like what does Colorado or California their ADMT rules require looking at the EU AI Act okay what does that require? And then trying to figure out what's the commonality there and how can we find kind of a reasonable path forward. there's no silver bullet, but something that we saw a lot kind of when privacy law was really kicking in in the early GDPR days and then to the US days where clients were kind of redoing efforts, right? They'd really focus on and hone in on one law and then that law would change or a new law would apply to them, and they just wasted a lot of can candidly time and money trying to comply with a lot of different things without waiting for it to be settled. I think that's really colored how we've looked at AI regulation and it's like, what is some commonality that we can look at through all the laws? Okay, it looks like we're gonna need some sort of risk assessment. Looks like we're gonna need some sort of transparency to, individuals or users. And very high level things like that where we can say these are things we can put in our a kind of AI governance program and we think, what if a new law applies to us tomorrow or no law ever applies to us, but we want to just be able to show racefulness we've kind of done our diligence there and we've taken just the key, the best concepts from a lot of these different regimes and put it into our governance program.

Gerard: Sounds good. I I like the the point you make about how there's it can be a difference between deployers and developers, but those that that's changing a lot that there's more and more. I think as this technology advances and and makes it easier to develop gives everyone kind of the opportunity there. we keep hearing the regulators are applying existing frameworks, consumer protection, data privacy. You were mentioning that, existing laws apply to AI. Uh and rather than waiting for for new statutes that are specific to AI, what should companies prepare for, for that kind of enforcement?

Tyler: Yeah, I think it's a big good news, bad news scenario. And I can just I think privacy is a great example and you mentioned it. So one thing that we really see whenever we're dealing with a privacy regulatory investigation, I think the the bad news side is a lot of times it starts with something very small and then it blossoms into something big, right? Because the regulator will come in and say, show me everything. And then I think that's something we have to worry about in the AI space, right? So A, it starts out with a privacy thing, privacy issue, but then all of a sudden now the regulator is looking at your AI. Or vice versa, right? Like, hey, AI's big and splashy, so that's what's gonna get the regulator involved. But now there's a deep dive into your whole compliance program. So I think that's the bad news is now this is kind of another vector to get that regulatory attention, which can then spiral out into into really a fine-toothed comb over your whole practices. I think the good news is is we've seen, especially in the US, but some in the EU, but especially in the US, we've seen regulators be, very willing to look at that riskfulness again. Right. And some just to keep using privacy as an example, some privacy laws in the U in the US even state like, hey, we'll take into account your reasonable efforts to comply with the law, even if we think you're not quite there. And I think that can be a little bit what we just talked about, like a guiding light for your AI governance program too, of okay, cool, how can we show that we put some real effort behind it? and of course, it's gotta be valuable, you have to be doing the right things, but on some level know what you're doing and the way that you're building your AI governance program is somewhat less important than the fact that you are building an AI governance program. So while you're taking that bad news of look now this is just another thing that a regulator regulator could hit on and it could spiral out into huge penalties and a huge investigation. I think the good news is is we've seen the willingness to you know say, hey, this, Company X is really trying to do this right. We understand it's a gray area, but they're trying to do the right thing and put in place some comprehensive governance. And I think that's something that, you I've been telling clients they should really take to heart is don't be let don't let perfect be the enemy of the good and get paralyzed. Like let's just start on some AI governance and put some decent things in place so you have some protection.

Nick: Yeah, that's a good point. And look, there's been a lot of talk around I guess risk tiered approach. So where consumer apps and safety critical systems, large scale models, or they all get treated differently. is that where you see regulation heading?

Tyler: Yeah, I think it has to. I mean, we've seen that, it's pretty much the only thing in the US that can get regulation passed. Colorado's a great example where it started really comprehensive and then, you very long story short, now we're at kind of a more niche target and it's kind of the higher sensitivity uses of AI. And, I just think it's it's kind of funny when you talk about when people say, Well, let's just govern AI, let's make one law that governs AI. being hyperbolic, but it's a little bit like, you somebody in the dot com era saying, well let's make one law that covers the internet, right? Like it's easy to say, quite difficult to do. And and I think it's just common sense too, of just, hey, people maybe are not as worried about some uses of AI, but there's more there's use cases that maybe we should be more worried about. And the law should reflect that, right? And let's let's put restrictions on the things that you are more dangerous. So I I think it's the only reasonable path forward at this point for sure.

Nick: I think you've you've somewhat covered this in what you previously said, but there's obviously a a fair amount that's already shaping the practice ahead of formal rules on these topics like procurement demands, risk registers, red teaming evidence. what do you think that in-house teams should be doing right now?

Tyler: Yeah, I you're I mean you're right. It goes to some of the flavor that we talked about. I just think, embrace it, start addressing it, right? I we have a lot of clients still that have a lot of paralysis around this and well, how do we do any internal governance or how do we put any controls on it? Or you could be the opposite side, right? Like how do we embrace AI at all because there's so much unknown and whether whatever side of that uncertainty you're on, I think it's, totally the wrong place to be. And instead embrace the use cases of AI in a reasonable way and then embrace the governance of AI in a reasonable way as well. I mean there's a lot we can be doing right now that is oh quite quite frankly like very low cost and very easy. But that, if something happens in the next six months, 18 months, three years, whatever, we can show like, hey, you know, Company X came to the table and they put a lot of time and thought into this. They have some controls. And you know, that's just gonna be a night and day difference no matter who you're talking to, right? Whether it's a a court, a regulator or or you court of public opinion. So I just think it's it's time to embrace it. Don't think you have to be perfect right off the bat, but start putting in some efforts. With that, I think it's kind of a good pivot. Nick, you asked me that last question, but I'm really interested to hear more about the entertainment industry. Obviously, you that's your cup of tea. I just really want to hear about, you obviously AI runs on creativity, but so does that industry. From your clients in that space, what question do you hear most right now from studios and talent about AI? Like what's your common issue that you're getting?

Nick: Honestly, it's sort of like two versions of the same question. So studios and labels are asking, can we use this? Um, do we own what comes out? And the talent is asking, can they use me or this, my performance without asking or without giving me some money or without my consent? So for studios and labels, you know, this is a a trade of title concern. Suppose a a a tool was trained on unlicensed material or the output can't be protected, that obviously causes problems for for financing, for insurance, for distribution. And nobody wants to build a franchise on an asset that they can't fully own, right? Or don't know the full sort of history of. So for talent it's about identity. So whether it's a voice or a performance style, and now it can be really quite convincingly reproduced in a matter of minutes. So obviously performers want to know where the consent sits, how they get paid and whether they can say no ultimately. So the good news is that I guess the conversation is somewhat matured. a couple of years ago it was I guess just panic. And now it's it's really the focus is on contracts. Yeah. so approvals, licenses, scope of of of rights grants and that's where I think the real work is is happening, at least on from what I'm seeing.

Gerard: And that answer you're talking about how the big concern can be, training on unlicensed data and aside from the the contractual controls, uh it's always interesting hearing, you the evolution of the fair use and the analysis of whether, training on on unlicensed data, you can be okay in some contexts or depending how it's used or use cases or or how the training happens. so with that, i is using copyrighted data to train AI fair use. What are what are you seeing?

Nick: I mean, well, first of all, it's a it's a really critical point because it goes to the heart of the creative business model. These creative industries make money by controlling copies of the works that they've invested huge amounts of money in and they license them. And AI training requires copies, like huge amounts of copies on an enormous scale and often without asking for licenses or without paying for them. So the question about whether or not something is fair use, I mean you would know better than than anyone. In the US, the answer turns broadly on whether that use is fair use or whether the the training is fair use. And as we've said before this is predominantly a case by case test. And the early rulings point in different directions. You know you've got competing claims going on right now. I mean, you had from last year you had the Barts and Anthropic judgment. The judge said that, training on books was fair use, but building a permanent library of pirated copies wasn't. that case resulted in a $1.5 billion class settlement. And then a related case, Cadvier versus Meta, Meta won, but again, only because the authors failed to prove a market harm. And it was an expressly narrow ruling. So this question of fair use is is very likely to be litigated for years to come, I suspect. I really don't see there being a a one-size-fits-all outcome, absent some kind of Supreme Court ruling or intervention by the legislature. It is just too highly fact specific for most cases.

Gerard: Exactly my view on it too. So that that sounds good. Let me ask you, beyond the contract and copyright issues, synthetic voices, digital doubles, deep fakes, you know, anything where it's impersonating a person, it raises urgent concerns about consent, compensation control. Uh how is the the growing patchwork of right of publicity and similar laws and regulations, evolving and i is it keeping up with these concerns?

Nick: I mean I think the short answer is my view is no, it's not keeping up. you know, again, the treatment, existence, and protection of publicity rights or name image likeness rights varies varies pretty significantly between jurisdictions. So if we take the US, for example, it even varies very significantly within the US itself. So the US has no federal rights publicity, it's state by state. The states are all moving at different speeds and slightly different directions. You got the Tennessee Elvis Act covering voice. You got California, who have rules on digital replica contracts and deceased performers. New York has added disclosure rules for synthetic performers and some sort of extended post-mortem rules. But you know, there have been moves towards a potential federal answer to this in the form of the No Fakes Act, but As far as I can tell, and let me know if I'm misrepresenting this, but there's still a long way to go for that to become law. And in the meantime, it's still pretty wild west. just to give you a sense of what it's like elsewhere, in the UK the gap is even sort of starker. English law doesn't have any kind of standalone image right or personality right at all. Performers basically rely on what's known as passing off, which is almost like a form of um unregistered trademark law, and performers' rights, which, again, they're not they're not tools that were built for AI generated images and deep fakes in mind. Earlier this year, the UK government promised a consultation on on this matter on publicity rights, but it still hasn't confirmed when that's gonna happen. So, in practice for now, again, what I'm seeing is these these issues are broadly being addressed through contract, negotiating payments, consent, approval rights, etcetera, because there is currently and probably for the short to medium term, a fairly significant gap in sort of the coverage from statute.

Tyler: So Nick, I think maybe just keeping the conversation across the pond, right, with your expertise. I mean, thinking about you the UK's position with copyright and AI training and and even opt out, it it seems like you think the landscape is quite different over there than here in the US. But I'm just really interested to get your take on, you how different you think it is and and what some of those differences are, just to keep elaborating on.

Nick: Yeah, I mean it's very different. I think the easiest way to describe it in very broad strokes is uh in the US we're seeing litigation, in Europe we're seeing regulation, and in the UK we're sort of seeing let's wait and see. the UK is no equivalent to the fair use exception. Our our sort of our copyright exceptions are very narrow. We have a version of the text and data mining exception, but it only covers non-commercial research, which wouldn't really extend to the types of AI-generated tools that we are we all know and use. the UK government did float an EU opt-out style mechanism where AI developers could train freely unless a rights holder had objected. But the creative industry's pushed back pretty hard. And in March, the government basically said that the opt-out was no longer its preferred option. And instead, like I said, it's taking a wait-and-see approach, let licensing develop, see what other countries are doing, and then decide the best approach. You know, Europe has the opt-out, and of course, the AI Act transparency rules. the German courts have been pretty bold, right, in their approach to this question of AI licensing and AI training. I mean Munich ruled against OpenAI on the lyrics matter and against SUNO as well more recently, even though SUNO trained predominantly in the US. So extremely interesting cross-border questions are starting to arise from this. And I think it's in it's kind of impossible to really to find a through line that marries all of these different jurisdictions. I think clients are gonna have to consider which is the strictest regime and and work back from that. And I think moving on then to some of the final points of this episode, I think here's the big one for you guys. AI is obviously moving faster than the law can keep up. And I think it's good for us to consider. Do you think we should regulate now? Do you think premature guardrails risk sort of chilling the beneficial uses of AI? So I'd love to hear what you think, Gerard and then Tyler.

Gerard: Sure. I I'll say kind of y yes to all of those. I I think we should regulate now, and I think we are doing a probably a pretty good job of regulating now. I think it's a balance that we need to be careful not to over regulate. I mean, I think, people are having a hard time understanding all of the nuances of AI and we're we're seeing how it's being adopted more. And I think passing, you extensive regulation could have too much of a chilling effect and could could be too problematic and of course difficult to comply with and all that too. But it seems like, you in the US there's been state legislation and a lot more has been proposed and been discussed than has actually been enacted. But I think that's a probably a good way to go about it, of thinking through the issues and not jumping in too far. And to one point Tyler made early on, there's a lot of regulation that applies to AI, even if it's not AI specific. And so, you know, having all the different federal and state agencies apply the laws to AI, I think, gets us a lot of the way there. There are some AI-specific things that probably should be regulated, but I I kind of think they're moving at a good pace of not rushing too much, but also not just letting the industry self-

Tyler: Yeah, and I'll counter it and I'll say, no, we should not be regulating right now. I think it's too premature. And Gerard and I have squared off on this in the past. I'm gonna be, a little bit overly hyperbolic for the sake of discussion. But no, I think we we've talked a lot about today, right? What is the risk and how it's already governed. Gerard was just saying it. So, no, I don't think I don't think the portrayals of, hey, it's the Wild West and anybody can do what they want with AI is accurate at all. And in fact, what I see is companies are very, very worried about AI and already putting their own kind of t constraints on it. And I'm just to be honest, I'm so colored by what we saw with the evolution of privacy law where, so many pr comprehensive privacy laws were enacted. And I think, you I'll be extremely generous to regulators and say I think only about twenty percent of those were actually good for individuals. and the most of it is just kind of additional regulatory burden. And, you know, the big winner there, candidly, is is people like myself. Like attorneys that can help big companies navigate these complex spaces. So yeah, I'm I'm definitely opposed to further regulation, but Nick, I guess I can throw it to you and you can be the tiebreaker. It's a lot of pressure.

Nick: You know what, I'm gonna I'm gonna go on and be consistent with the the British government and sit on the fence and say, I don't think we know yet. I'm gonna wait and see.

Gerard: Yeah.

Nick: I just yeah, I wanna see what the other countries come up with and the results that come up as a as a result of that. And I think that's probably a nice place for us to finish this first episode. going forward, the three of us will keep on hosting and bringing in colleagues from across Reed Smith and beyond to tackle the AI questions that cut across all of our respective worlds. And look, if there's an AI issue that you want us to take on, then we want to hear about it.

Outro: Tech Law Talks is a Reed Smith production. Our producers are Shannon Ryan, Amanda Saunders, and Mason Kautz. For more information about Reed Smith's Emerging Technologies Practice, please email [email protected]. You can find our podcast on all streaming platforms, reedsmith.com and our social media accounts at Reed Smith LLP.

Disclaimer: This podcast is provided for educational purposes. It does not constitute legal advice and is not intended to establish an attorney-client relationship, nor is it intended to suggest or establish standards of care applicable to particular lawyers in any given situation. Prior results do not guarantee a similar outcome. Any views, opinions, or comments made by any external guest speaker are not to be attributed to Reed Smith LLP or its individual lawyers. 

All rights reserved.

Transcript is auto-generated.