In this episode of The Good Bot, Brett Mason is joined by Rusty Close, Austin Padgett, and Brent Hoard to explore the complexities of contracting for artificial intelligence (AI) in health care.
In this episode of The Good Bot, Brett Mason is joined by Rusty Close, Austin Padgett, and Brent Hoard to explore the complexities of contracting for artificial intelligence (AI) in health care. The conversation points to key intellectual property concerns, including ownership of AI-generated content and the impact of recent Copyright Office guidance.
THE GOOD BOT s02e07:AI Contracts – IP and Privacy (No Infringement Intended x-over)
Recorded 7/23/25
Brett Mason: Welcome everyone to The Good Bot, a podcast focusing on the intersection of artificial intelligence, health care, and law. I'm Brett Mason, your host. As a partner and trial lawyer here at Troutman Pepper Locke, my primary focus is on litigating and trying cases for life sciences and health care companies. However, as a self-proclaimed tech enthusiast, I'm also deeply fascinated by the role of technology in advancing the health care industry.
Our mission with this podcast, is to equip to you with a comprehensive understanding of artificial intelligence technology, its current and potential future applications in health care, and the legal implications of integrating this technology into the health care sector.
I am excited to be joined by three of my fellow Troutman Pepper Locke partners today. The first is Rusty Close. Rusty is an experienced technology attorney, which is perfect for this podcast, and he routinely advises and counsels companies on complex, transactional, and intellectual property matters. So, Rusty, thanks for joining us.
Rusty Close: Glad to be here.
Brett Mason:Also joining us is Austin Padgett. Austin Padgett is also a partner in our Atlanta office of Troutman Pepper Locke. He is a copyright, trademark and advertising attorney in our intellectual property group. So, that is going to be a lot of the focus of what we'll talk about with Austin here today. Thanks for joining us, Austin.
Austin Padgett: Yes, the pleasure is mine.
Brett Mason: Finally, we are happy to welcome back to the podcast, Brent Hoard. Brent is also a partner at Troutman Pepper Locke, and he is in our privacy and cybersecurity and data privacy practice group, and we're excited to have him talk about the interaction of IP and privacy with us today. Thanks for joining us again, Brent.
Brent Hoard: Great to be back.
Brett Mason:So, let's just get started with the basics. Our goal today is to talk about intellectual property and privacy considerations, specifically in contracting and negotiations around these issues. So, Austin, why don't you start us off and explain what are some of the main concerns when it comes to artificial intelligence and intellectual property, negotiations, contracts, in a health care-type setting?
Austin Padgett: Yes. I put the high-level concerns in two categories. First, with pre-existing content that we own or we have a license to, what can the other party do with it when it comes to using or building AI-based tools? That's an issue of control over your own assets. So, that's the stuff we already own, the old stuff, if you will. The second category relates to new stuff. In the creation of new content is an AI solution being used, and if so we've raised a question then about the ownership of that new content.
Brett Mason:How does AI complicate the questions that you're asking about ownership?
Austin Padgett: Yes, let's take a case where a health care organization hires a marketing agency and that marketing agency is going to use some AI tools. If the AI creates content and we're still working out these issues in a set of cases that we're watching and/or participating in, but the Copyright Office itself has set out a white paper that tells us that such content that the AI creates is not copyrightable. That means the organization, and really no one, can own it in the traditional sense, which is a significant concern if you haven't anticipated that in your agreement or your business plans.
Brett Mason:So, you just mentioned that white paper by the Copyright Office. Can you tell us more about that and how you're taking that a consideration when you're advising your clients?
Austin Padgett: Over the past year, the Copyright Office has released this multi-part report on AI and copyrightability particularly. It doesn't carry the force of law per se, but it outlines the office's position on the issue. And the report emphasizes this constitutional requirement for human creation, for copyright protection. It references cases like the PETA Monkey Selfie Case, which is a relatively famous copyright case, to illustrate these principles.
So, if you're looking for that discussion in the white paper, you want to look at what's labeled as part two, labeled copyrightability.
Brett Mason:So, for our listeners, what is the PETA Monkey Selfie Case? Because I don't think most of our folks here will have heard of that.
Austin Padgett: Okay. This is a case where an artist went out into the jungle and put up a camera. It so happened that - we'll call it a monkey - there's this specific species that's referenced. I don't have it off the top of my head, comes and takes a picture of itself as in a selfie. Then PETA used this on behalf of the monkey to present a copyright action over the use of that photo and the distribution of it. And PETA would have its own purposes in that because this is the question of personhood. Does the monkey get a copyright in that photo the same as a person would have?
The court ruled that no, a monkey is not a human and therefore there can be no copyright. So, PETA's plans were thwarted for that issue on that day. But that's the nature of the case and that principle is the important one. Human involvement. The Constitution is where the copyright clause is embedded with the patent clause as well. Ultimately, everything goes back to that, and the copyright office is deriving and trying to explain that in its view, copyright requires human creation.
Brett Mason: So, in this instance, AI-created content is the same thing as a monkey creating content.
Austin Padgett: For our purposes, for sure, yes.
Brett Mason:Okay. So, that kind of covers the high-level issue of the second category or the second bucket you mentioned, but let's go back to the first category, which would be the pre-existing materials that a company already has. What are the concerns there and what advice are you giving to clients when they're contracting around that?
Austin Padgett: What we have is a series of ongoing cases right now that relate to how existing material interacts with AI. So, if you have a batch of content, can I take it and build my own AI model with it? If I'm providing services to you, can I plug your content into a third-party AI tool, let's say, to help me build my analysis or whatever I'm going to do for you as a vendor or something like that.
The courts are considering whether these types of uses are fair uses that are not subject to copyright prohibitions. Well, if that's the case, that's tough for the content owner that didn't anticipate that kind of use and agreement before handing over the company's most important work product. So, if you own content, you typically want to control its use and know what’s happening with it.
In agreement work, I don't want the default rules of fair use or any of the other default rules of copyright ownership or copyrightability to apply. So, I think about where can I set out specifically the party's respective obligations. One issue that we've had, typically in an agreement, you'll see that we've always assumed that there was human involvement and creation in these types of works. So, we've always assumed copyrightability as well. You’ll have reps and warranties about, okay, you're going to own this or you already own this and you're giving it to us, those sorts of things. Now, what we see are additional representations about the actual ownability of the stuff that's being made or that's being licensed.
Brett Mason:What I really appreciate about what you just said there is not just relying on kind of the default, the default that's been used prior and what we have expected along the way. I think, Rusty, that really ties us into what we've talked about in preparing for today. How does that affect your negotiations and your contracts? We don't want to just use the default, right? We want to actually specifically lay out what we intend to do with these various issues.
Rusty Close: Yes. I mean, one of the big things Austin and I always talk about is don't rely on the defaults that intellectual property law provides for us, because it just didn't anticipate all of the things we deal with in this day and age. It's just not set up that way. So, what we're always counseling people on is, I always use this term, like be really intentional in your use of these tools. So, that can be the way you integrate them into your products, the way that you use them to help accelerate the development of your products, all of those kinds of things. Then, it's like, well, what do you mean when you say be intentional? Be intentional. I think it's understanding the implications of their use and how incorporating them into your products might have an implication. Like Austin said, some of this stuff is unsettled.
So, are you doing something that on the one hand is making you more efficient, making your development process faster, whatever the case may be, but on the other hand is going to potentially cause problems for you down the road.
Brett Mason:One of the things we talked about, again, being intentional in the contracting is how important it is to set clear terms in the contract so that both parties understand what they mean and to hopefully avoid future disputes. So, can you tell us a little bit about that and some of the terms that you're having to set around artificial intelligence?
Rusty Close: If we envision a scenario where you have a product, you've got a software tool, a SaaS platform, whatever it might be, and let's just say that it does incorporate some aspects of AI. Maybe it's in the way that it analyzes information that your customers put into the system, whatever it is. It goes back to what Austin was saying, but if you're the customer in that situation, you want assurances that the software provider is using those tools in the right way, just to kind of put it at a high level. So, what you'll see is any use of AI tools, just generally speaking, is going to conform with AI law. Then we'll define AI law as best we can, referring to EU Artificial Intelligence Act or different types of regulations and things of that nature.
Again, some of it's going to be unknown, but we want to put the onus on the people who are making these tools available in this situation, that they're using them the right way, that they're not going to take our information that we're putting into the system that's important to us and let it be distributed far and wide or in a way that we can never get back.
Brett Mason:Yes. I think defining AI law might be one of the most difficult parts of that because as we've talked about a lot in this podcast, it just keeps changing, and we know that the law is always pretty slow to catch up to technology. So, good luck to you on doing that.
Do you recommend defining Generative AI in these contracts? And how do you define that?
Rusty Close: It's going to depend on the scenario. But where this might come up is, let's say that you've got a software tool and someone's doing diligence on your company because they want to make an investment or an acquisition, we see this come up in the diligence ask, or they're asking the target to rep and warrant whether they have or haven't used Generative AI in the development of their product. Yes, you'll see a definition. It can mean a lot of things, but it's a specific type of AI that's focused on creating models or generating new and original content or whatever that might be. But you will definitely see that, especially it's getting more common in these, like I said, in these asks and these diligence requests. It goes back to that same question of, are you being intentional in the way that you've used these tools? Are you understanding the implication?
Like Austin said, they're trying to settle, okay, well, imagine a scenario where you're using AI to create aspects of your code base, but the Copyright Office is telling us that, well, that may not be copyrightable. So then, when you're repping and warranting to a potential investor or acquirer, yes, we own everything that's in our code base. Do you actually own it? Do you truly understand whether or not you own all of those aspects of your code if you were using Generative AI to help create some of it?
Brett Mason:Anything else that you would recommend for companies when they're being intentional around these issues?
Rusty Close: I hate to keep harping on that word, but it really goes back to just being intentional and having processes in place, educating the people in your organization, or your developers, or the external developers that you're working with, understanding how that those folks are going about creating your platform and all of the things that they've done that have gone into it. Because I think things can happen quickly without your truly understanding how things came about if you're not keeping a close enough eye on it.
Brett Mason:I think one of the things you mentioned, Rusty, when we were preparing for today that I thought was interesting is that you can't just go about business as usual with these types of issues, and you can't just try to shoehorn dealing with these issues into current IP law or current copyright law. Perhaps, part of your advice that you're giving to your clients is to make sure that all of their employees or folks who are engaged in this area are not doing that, that they're not just going about business as usual, assuming everything is going to run the same way. Is that fair?
Rusty Close: It's definitely fair. I mean, I think there is a real appetite for using these tools because they seemingly can make life much easier. I think we all know that if we can take a shortcut, we might want to take that shortcut. But it's really important to understand, to have those policies in place, whether it is don't use them, we just don't know enough about them at this point, or use them, but you have to check your own work, you have to create your own work based on what you get out of it, whatever that might be. But it's not just turn people loose and say, “Hey, use it. Whatever happens, happens.” It's definitely not that.
Brett Mason:Yes. And I think we've seen a change in companies maybe at the very beginning, they were saying don't use it, and now they've realized that that's not really tenable and they have to. So, they're moving towards, again, like you're saying, being intentional around use.
Rusty Close: Hopefully.
Brett Mason:This connects a lot with Brent, things that we've talked about in the privacy space. So, give us a perspective of how does privacy tie into this again in these negotiations and these contracts? We're worrying about the intellectual property, worrying about the copyright issues. Now, let's layer on privacy to that. How does that work and what are you saying?
Brent Hoard: Yes, I think privacy in a way is running in parallel with IP and it has similar concepts. We're looking at the two different buckets. There's the input into the AI on one side, and then also the output, and what's coming out, what are we doing, and what's the use case? All of this involving personal information in a health care setting, potentially regulated in personal information. I think that, to Rusty's point, and being intentional also applies on the privacy side of the house and that's really around the data and understanding what data is going into the model and how you're using it, and then what type of personal information related decision making potentially is being run through the AI, and where is that data going when you're doing the processing.
Brett Mason:So, what are some areas of focus for privacy when you're analyzing AI issues in a potential transaction?
Brent Hoard: Breaking it into those two buckets, the first thing we're really looking at is understanding data rights. So, if we're training an AI model and you get past the gate that we're using some kind of personal information, we have privacy in play. Is there a right to use that information to train the model? That's looking at a privacy policy, and potentially contracts with customers, if it's customer data. If it's HIPAA, is it de-identified or is it PHI? There are potentially more restrictions in play there. Also, looking at other data sources, there are a lot of data brokers or other resources that will provide big pools of information that you can use for training purposes. If that is the case, you really want to look closely at the rights and make sure that you're not training an AI model on data that you shouldn't be using to train that AI model. It can completely foul everything up and you end up with a model that you can't use or you might even have to destroy.
So, that's on the front end. I think in terms of output and understanding the third party that's processing it. Is this an in-house on-prem solution that you're using or do you have a third-party tool? If you have a third party involved, there's data sharing. So, in this context, it's personal information. I usually would say at least a data protection agreement or something to understand what's going on with that data. If it's PHI especially, you definitely need to have a business associate agreement in place. When you have those third parties, to Rusty's point, it's out in the world, potentially. So, you really need to be cognizant of any type of identifiable information and where it's going.
The final piece to think about is a lot of the emerging laws. Outside of HIPAA, there are various state laws that could come into play like Washington's My Health, My Data Act. Virginia has a new law around reproductive health information, which really requires consent for any kind of data uses and captures businesses that are not traditional health care that might sell products that are captured or even provide services to companies that sell products that are covered. New York has a forthcoming health information privacy law, which also has consent requirements. So, it's really understanding, I think, if I were to break it down, understand your data flows and what you want to do with that data. If you can figure those pieces out, you can put in your organizational controls and things that you would need to make sure that you're handling your approach to AI properly. If somebody's coming in to look for a transaction or you're entering into contracts with customers, et cetera, you have to have everything buttoned up.
Brett Mason:Well, thanks for that insight, Brent. I think that's really helpful to layer in with what we've heard from Rusty and Austin. So, as we wrap up, just for anyone who's joined us, what final advice would you give to businesses navigating these issues?
Rusty Close: I think it's just don't take it lightly. In some cases, it seems too good to be true. It seems too helpful to be true. It seems like we can cut down on months and months of work in a really short time. That might be true, but again, you have to understand the potential implications.
Brett Mason:Well, thanks, Rusty, Austin, and Brent for joining us. I appreciate your insights on things people need to be thinking about in IP, copyright, and privacy when they're doing or transactions or contracting around AI. So, thanks again to our listeners. Please don't hesitate to reach out to me at brett.mason@troutman.com with questions, comments, or any topic suggestions. You can also subscribe and listen to other Troutman Pepper Locke podcasts, including an intellectual property-focused podcast starring Rusty and Austin, wherever you listen to your podcasts, including on Apple, Google, and Spotify.
Copyright, Troutman Pepper Locke LLP. These recorded materials are designed for educational purposes only. This podcast is not legal advice and does not create an attorney-client relationship.The views and opinions expressed in this podcast are solely those of the individual participants. Troutman does not make any representations or warranties, express or implied, regarding the contents of this podcast. Information on previous case results does not guarantee a similar future result. Users of this podcast may save and use the podcast only for personal or other non-commercial, educational purposes. No other use, including, without limitation, reproduction, retransmission or editing of this podcast may be made without the prior written permission of Troutman Pepper Locke. If you have any questions, please contact us at troutman.com.