EP175: The Fine Print: AI, Marketing And Brand Risk

Speaker A: Hello and welcome to another episode of Leader Generation, brought to you by Mattup. I'm your host, Tessa Berg, and today I'm joined by Candice Kirsch. She's a partner at Frankfurt Kernet Klein and sells in the advertising and marketing group. She has received countless accolades and awards over the years in the field of advertising and law. And today we are going to jump into that intersection specifically really starting to better understand where AI marketing, intellectual property, procurement, agency operations, and all the things that you're stressed about as a marketing leader today. Reading contracts, trying to better understand what you need to worry about when it comes to generative AI data and who owns what. So thank you so much, Candace, for joining us. I'm very excited about this conversation.

Speaker B: Thank you. I'm excited as well. I really enjoy talking about all of this, so this is going to be great.

Speaker A: I know you have some speaking engagements coming up, but before we start, tell us a little bit about your background and what is coming up next for you.

Speaker B: Thank you. Yes. So I am, as you said, in the advertising group of Frankfurt Kernit. I have been an advertising lawyer for many years and I've had the pleasure of working with some of the most amazing, innovative and creative agencies over the years, as well as some of the country's top advertisers. And through it all, working with creative people, it just keeps you constantly on your toes and applying all different areas of law. It's never boring. It's really interesting. And so, yes, I do speak at various places around the country on currently a lot on AI. I have a couple of speaking engagements coming up, as you said, and I'm, um, uh, just looking forward to this conversation.

Speaker A: So one of the questions we get a lot is, are people keeping up with the pace of AI and what do they really need to focus on? So I think it's no secret that AI is moving at breakneck speed, that policy and regulation is not keeping up. What should people be worried about right now and what should they really be focused on when it comes to their own use of, of AI in marketing and advertising?

Speaker B: That's interesting. I mean, I think people right now, as you said, everybody wants to jump on the AI bandwagon and I do think people think it'll make things quicker, faster, efficient. However, to some extent that is true, but you do have to take a little bit of a step back because there are a lot of legal issues and pitfalls and things that can come up that really can cause a lot of issues for you as you're embarking on your advertising efforts. And I think it's really important for people to understand the issues and therefore to think about it and to put in place protections to protect the agency or the brand from potential liability as a result of this. You know, jump first and think about it later.

Speaker A: And what are some of the things that you see that are pretty common in protecting agencies from liability?

Speaker B: Well, the first thing is you have to think about how you're using AI. Um, so AI can be used in concepting, ideation, AI can be used in research, AI can be used in uh, creative and execution. And so, and each one has its own types of issues attended to it. So for example, if you're going to use AI for um, claim development, you've got to be worried about the AI, just the, also the input and the output. So the input, again depending on the type of AI, taking even a further step back is, you know, are you using enterprise licenses? What were, you know, does your firm have a policy? Should you be using closed enterprise licenses? Is it worth the cost? Because again, trying to save money with publicly available licenses, your input can be used for training purposes. There's confidentiality concerns which your people should be aware of, and then you have to be worried about the output and hallucinations. There's cases out there where, you know, particularly in advertising, the fact that your claims, uh, may have been developed by IP are not going to give you the protection against false advertising claims. So you do need to consider, even in claim development and how your use of AI, you know, the fact that AI says it, and I think most people know that at this point doesn't necessarily make it true. And you as a company are still responsible for the, for the claims that come out.

Speaker A: You mentioned that companies are trying to save money by using some of the free models or they might tell people, hey, we really need you to use the licenses. We have a chatgpt. But perhaps someone really likes Perplexity or Claude or a different model and a lot of companies don't realize that opens them up to risk.

Speaker B: Absolutely. And the benefit of an enterprise license is there are going to be, depending on what you're using and which AI you're using, there may be prohibitions on use commercially. There certainly there's the possibility of using your information for training. Uh, you know, and if you're not really looking at that, I mean, it's one thing if you want to create an illustration in your PowerPoint presentation of Boy with Cat. Yes, it doesn't really matter if you use a publicly available Sourced AI. But as a company, I think you need to put policies and procedures in place. And even if, yes, of course you're not going to hinder somebody creatively, but somebody should then be reviewing those terms and conditions and making sure that they are appropriate for the use. Because again, what each of your individual employees could be doing could be imposing liability on the company. And as a leader of a company, you should be aware of that and you should be controlling for that. And so again, I come down to education and policies, whatever they may be, are such a great thing for a company to have. I mean, it really is so important because I don't think people understand the potential, um, damages or liability that they could be exposing the company to. And so by training like this kind of training or showing up like when I do these sessions, really explaining to people what the pitfalls are and how you can easily avoid them is just, I think, gives the company so much more protection.

Speaker A: Yeah. And I think a lot of companies or people who are listening to this podcast right now may be thinking, I've tried to tell the team that there's liability and I know there's still, you know, dark AI or dark AI use happening, but external voices carry a lot more weight. And I would just encourage like marketing leaders, like sometimes you need that external speaker to come in with really sound use cases and examples from where they've seen other businesses. Because just like the Internet, I think that there's still this. When the Internet first came out, when people weren't highly aware of the consequences of sharing their own personal information. And now that's pretty well understood that if I am entering something into a form, I better make sure that that form is, is going to a company and that payment security is incredibly important. But a lot of people had, uh, to go through a lot of pain to learn that. And AI accelerates the risk. And so the more you can have external voices and people really understanding the risk, they're putting the company at where the liability sits when they use unlicensed or the non enterprise licensed software, the better.

Speaker B: Yeah, it's interesting because, I mean, I grew up with creative agencies and I really have had the pleasure of working very, very closely with them. And so a lot of lawyers, at least reputationally can just say, no, this is what you can't do. The goal here is not to tell everybody what they can't do and that if they use AI, you know, the entire world's going to fall apart. The goal is to try to find ways to do what they want. To do in a manner that protects the company and that makes the difference. Right. So if you're bringing in training just to say you can't do anything, you know, go back drawing stick figures and you know you're not going to use AI for any. That's an unrealistic view. And, and I think as you're looking, we're going to talk later and I don't want to jump ahead, but just a little bit for a moment is contracts. When the contracts say you can't use AI or there's no use, these are unrealistic. I mean everything has AI at this point. A little exaggeration but embedded in it. So it's really a question of understanding the AI and understanding what your corporate risk tolerance is and then putting in place programs that allow people to do what they want to do, but do it in a way that doesn't expose the company to liability that it's not even aware of.

Speaker A: Mm mhm. Yeah. And that brings us to like our next set of questions. So as an enterprise or a business, I've decided to license an AI enabled service or a platform or app or piece of technology. And now I'm looking at the vendor agreement. It gets tricky again. And where do you see companies unknowingly taking risk when they do license and deploy these AI tools in the business?

Speaker B: So the first thing, I mean it's kind of walking through, you know, when you're using AI, what type of AI and what's the purpose of the AI? Because as we were talking about before, you know, I think one of the big issues that people are starting to become aware of is the ownership. Even if as between you and the platform, the platform is not going to own your material from a copyright perspective. And a copyright gives somebody the exclusive right to use creative material that's fixed in a tangible form. So. Right. So you have certain protections. You can prevent others from using your creative if you own the copyright. If you don't own the copyright, if something is free to be used, anybody can use it. The problem with AI and the Copyright office has made clear, at least right now, is copyright in the, in the US requires human authorship. And because AI, uh, the authorship and the creative decisions are determined algorithmically, you can't get copyright protection in something that is wholly created by AI. So from a company perspective, you know, it's like, oh, I have to own everything. Well, the question is, as you're looking at it and as you're thinking about this issue is does it matter? Again, this is that Whole balancing, does it matter in the. For what you're doing? Right. In other words, like if I'm just doing some quick hit social media stuff, does it matter if I don't own it? If I'm just taking my product and putting it 100 different scenes, does it matter if I don't own it? If I'm just doing a scene extension of a brick wall, does it matter if I don't own it? And then as you delve deeper, and this is again the importance of training, as you delve deeper, there are things that you can own. And then it also may make the ownership of the AI component less important. For example, if there's a case in which somebody created a comic book and they use midjourney to, I believe, to create the illustrations, well, in that situation, the person, you can own the copyright in the text, you can own the copyright in the arrangement. And so does it matter as much that you don't own the copyright in the individual drawings? If you are basing your AI output on something that you already own the copyright in, and it's just a derivation of that, there may be elements of that derivation that you don't own, but you're going to own the copyright and the underlying input, the photograph, whatever. So does it matter again? So that's why it's such a nuanced situation. Because if it doesn't matter, then this whole notion of not having ownership, you know, again, is not a barrier to what you're doing. On the other hand, if you're creating a spot and you have to own it and it's really important, you want to prevent others from using it and blah, blah, blah, then maybe it is. So understanding leads to intelligent choices.

Speaker A: I think that's a really important call out and one that truly I didn't fully understand. So I'm glad we're talking about it. But that's an important conversation to have with your clients. Because of course there are clients who, you know, have really high value IP and you know, characters, toys that they've developed that they're monetizing in many different ways and merchandising and commercial and other, and they do have in their contracts. You can definitely never use AI because we won't own it. And that's number one thing cited, even this conversation, really. And I think that the thought of not owning something, it's just scary. But when you break it down, it's like we can prioritize what's really important to own. And the core of our business, our characters are Our core ip. Yes, that's still important to own and we'll own that. But if, would I still own the outputs if all I was doing was using it to translate it into like different languages?

Speaker B: Well, you would own the output, you would own the copyright in the text itself that you wrote. So does it mat, I mean uh, you're talking about again, does it matter that you would own the underlying. So somebody couldn't just translate it into another language and you know, be able to still use it by virtue of the fact that it's still based premise, it's derivative of your work. So again if it's a slavish copy of something on which you own the copyright, I mean there's nothing really to own at that point or not own because the, the underlying copyright in the work is yours. And it's interesting. And then of course the other thing is from an infringement point of view, what people need to understand also is that you can create the infringement through your input. And this is this whole notion of policies of recording prompts and making sure that your people are not inputting something that says I want you to create um, a photograph, you know, and put in somebody else's photograph that you don't own. And then you get something derivative. And. Or I think one of the newspapers did, you know, I don't know that you could still do it, but did a um, experiment where they were saying things like put in sponge under the sea with pants and they got spongebob squarepants. So I mean if you were to, you know, you can try to force the AI and that, that, you know, the fact that it's created by AI doesn't mean you can't infringe on a third party. Right now most of the lawsuits that we're dealing with are based on the creation of the AI itself. There's a lot of different lawsuits, whether it's on infringing on the copyright in something or voices, et cetera. But um, you can create that infringement by how you input it. And so even though currently with I am not, I'm um, just trying to, off the top of my head, I can't think of a lawsuit that's based on the output alone that somebody inadvertently and perhaps it's statistically unlikely, it is possible. And certainly if you are forcing that infringement, it becomes a uh, much more likely scenario. So again, teaching your people not to use third party uh, input, that they don't have rights to, to create the AI and to keep prompts and understand what their input prompts are so that they can ever defend if you do get a claim is really important.

Speaker C: Marketing leaders AI isn't just the future, it's happening right now. But is your team ready? At modop, we don't just help you adopt AI, we help you transform the way you work. Our marketing AI audit evaluates your tech stack, responsible use policies and your biggest AI opportunities. In just six weeks, we'll build a roadmap for smarter automation, better insights and real efficiency gains, ensuring your team isn't just keeping up, but leading the way. Get started today at, um, modop.com AI that's M-O-O-P.com AI this has been very enlightening already.

Speaker A: And we didn't even get to the third area of what I think sort of blocks brands and creative folks from using AI and using it properly, which is data. A lot of people are concerned about. Are they exposing their data? What are the measures in safety and security to protect owned data from going out into the public or even their own inventions? What are some of the ways that people are managing that liability?

Speaker B: Well, I think using the closed systems and also having policies in place is a way certainly to manage. And also only using AI platforms that have check the data security, uh, of those platforms. It's a little bit off from data, but where I'm seeing a lot and where a lot of the laws are, and I think it's important for your listeners to understand is the issue which everybody's focused on now is digital replicas and synthetic performers, which I think is in the creative area, probably one of the more exciting and also has the most pitfalls, um, in terms of because everybody wants to take somebody and use their voice or their style and create, it's much cheaper as opposed to going into studio with a performer, it's much cheaper to give somebody, record them and then create AI versions of them. And so you're seeing a lot of laws now between sag, AFTRA and the States. That's really a hot area right now where all the laws are coming in to really regulate how people are doing that. And I don't know if we have time to go into it all, but it's certainly another area that, uh, I think is of tremendous interest to a lot of people in the creative field.

Speaker A: Yeah, now that one, it definitely comes up a lot. But just to ask a broad question on it, if I am, um, an actor, like a voice actor, is there an avenue where it's legal for me to allow someone to use my voice? So, like, I Don't have to go in and record. And I've given permission for you to use like the AI version of my voice to do work.

Speaker B: Yes. So you look at that. There's. When you look at digital replicas, the laws are. There's a couple laws that are currently out there and there are more that are currently. They haven't been passed yet, but they're certainly being contemplated. But the big ones right now are New York and California. And what they require is if you're taking away work from a talent and you need to make sure either you are reasonably specific as to what you're doing. So it's not just, you know, when we used to do talent agreements, we used to say any and all media, forever and ever, whatever, because you would just take. And they would do one record and you would. Now because of the possibilities, you know, the laws really want to make sure that you're either specific or the person is represented by a lawyer or a union. We now have a law in New York as well that deals with models. That doesn't really. The fact that you have a lawyer or a union is not going to be enough to avoid this reasonable specificity. The whole point is that you don't want somebody 20 years from now to still hear their voice and to feel like they were taken advantage of. And so they want reasonable specificity. What that is, is, is the question that lawyers are debating because, you know, how much specificity is enough. And again, those laws are only in certain states currently.

Speaker A: That is really interesting. The more interesting part of actually what you said is that it's only in certain states because it's kind of hard, if not possible, to run things just locally. Especially in the states that you're talking about, which have large populations. Do you typically advise clients to follow like the most strict policy and rule to protect themselves?

Speaker B: It depends. I mean, you know, that's, that's obviously the simplest way because you don't know what's going to happen down the road. And certainly, um, that is the simplest. But it depends on the situation and what you're doing and how you're, you're doing it. The problem in our country is that we don't have a federal right of publicity. These are all state specific and the state laws differ. So you either, you know, do go to the most conservative and then less to worry about, or you, you know, make a much more surgical decision as to how, what you're going to do based on, you know, where, where you're going to record what your usage Is, et cetera, because, you know, these things are. Can be very fact specific. And then you also have the issue of synthetic performers, which we thought, okay, fine, I'm just not going to even bother with people. I'm just going to create a, you know, whole cloth performer, you know, which is synthetic performer. But now we have laws which just went into effect in New York, for example, and there's three other states, I believe, that are contemplating similar laws or two other states contemplating similar laws that require labeling. So that if you use a synthetic performer, you have to let people know that this is not a real talent. And then you have, on top of it, you have the platform disclosure. So there's the whole other issue of which is another interesting area is do I have to disclose? And I think a lot of people are facing that, and that becomes also a quagmire. So, you know, there's a lot here. It's certainly possible, but. And it's a very interesting analysis through your campaign. The point is it's doable. It's just a question of, you know, going through it. Making your adjustments and doing it up front is frankly the best way. Because if you're able to come into a campaign where, you know, explored your creative. I mean, I work with my clients, they have a deck, we talk about it. And then you can make modification as opposed to finding out at the very end that you're going to have to do these things or you can't do these things. So, yeah, and I feel like you

Speaker A: were saying before, where you really need to focus on protection and managing risk is based on how you're using AI in the conversation of labeling, whether or not this is a synthetic performer. And going back and forth, I feel like you have to think about what would your audience think? You know, like, is your target audience going to be totally freaked out that this is a synthetic performer? Then maybe not only do you have to label it, but don't do it because you could land very poorly.

Speaker B: And then on top of it, you have like the other, you know, advertising issues, which is if it's a demonstration, if I have a beauty product and I'm showing somebody, uh, an AI synthetic model putting on that beauty product and showing how well it shows your eyes. I mean, is that you have to. Basically it's now a false demonstration. Depending on the context, if I have an endorsement, if I have fake, you know, bot saying, this is the best product I've ever used and I use a lot of products, you know, you're going to have to Let you're going to, just for truth and advertising purposes, you're going to have to label it. So it definitely makes my job interesting because as I said to you when I started this, the goal is how do I get the creatives to be able to do what they want to do safely and within the company's policies. And again, versus no AI I'm going to close my eyes. It just doesn't uh, work that way. And then to your earlier point is when you use production companies or third parties, how do you push that obligation down so that they are not using A.I. uh, in a way that makes you uncomfortable. So you, you know, you don't realize what they're doing, but it goes down the chain. So all of it is just, it is counterintuitive when you're thinking that AI is faster, quicker, easier. But it can be faster, quicker, easier. But you can't just go as fast as you would like. Like anything else, you have to give some thought to some of these other issues, in which case you can get a great campaign that, you know, uh, maybe faster, easier, you know, but if you end up falling into one of these pitfalls, you end up with a uh, liability that you had never intended, which makes things much more expensive.

Speaker A: Yeah, I mean I feel like it's a lot of things when you invest early in the overall strategy and then you put in the right guardrails, processes and to your point, training, education. So now we understand the reason for the strategy, the policy and the uh, through that training then you can move faster after that. But there is no like on, off, switch to now we're using it, now we're not.

Speaker B: Yeah, it is such a great tool. And you know where it is a bigger philosophical conversation as to whether it becomes another tool in the, in the toolbox or it becomes something, something bigger that I don't know. But it is something that is incredibly powerful and incredibly, it is real. It's going to be there and you have to use it. So the question just is how to do it. You know, as I said it's to me it's like anything in the creative industry. It Teamwork, you know. Right. Bringing in. Um, I'm very fortunate to have long term relationships with my clients and so I think there's a trust there. But bringing people, bringing in your lawyer, bringing in people to really sit back and understand what you're doing. Somebody who's going to bid out, you know, the specs with understanding what AI, uh, somebody's looking at those terms and conditions and Again, it may add an additional layer, but it's certainly still going to be faster. It's great in its proper place. Mhm.

Speaker A: Yes, I agree. So you've spent your career helping those long term clients navigate big shifts like this in marketing and advertising. When you think about everything that's going on now and how that might land in the next maybe two, three years, if you could, what do you think is going to be surprising?

Speaker B: Oh, that's a tough question. Ah, you know, all of this is going to be surprising. I don't know. As a consumer as well as a lawyer, I don't know where this AI is going to shake out. I don't know if we're going to have this dystopian view of there are no jobs, everything's going to go by the wayside, or we're going to have this be just another great tool to make life easier. Um, it's certainly, it's hard to tell. And um, I choose, as trying to be the perennial optimist, I choose to think it's just going to be a wonderful tool in the toolbox and that uh, this dystopian view is going to be left to Hollywood. It is interesting. I can never predict and it's. The wonderful thing about my job is that I have been fortunate enough to work on such great creative campaigns over the years and I've been able to work with some of the greatest creative minds in the business and I just uh, my job is to try to help figure, help them figure out what to do as opposed to trying to. Because I can't keep up with them in terms of their creative visions. So it's always. And that's what's fun. I never know.

Speaker A: Yeah, no, that is fun. I feel the same way. Like I choose to have a optimistic view and there's some data coming out now that I think validates that human creativity and human problem solving is going to be, continue to be in high demand. And it's the simple fact that companies who do try and find that least path of resistance to just license something, to just try something, just do something, aren't able to measure the ROI impact of that. And again just have to like whether it's with managing risk, managing liability, or trying to measure what something will look like at scale, you do, you have to invest, you have to invest real, real dollars. Yeah, real dollars. Real human power behind. How are you going to do this? Well, in a way that serves your customers and the people at your business.

Speaker B: I mean you're absolutely right. And it's going to have a place, but from what I'm told, I mean, certainly not there yet, where great creative still requires humans. You know, I can speak to it from products of, uh, you know, not to malign anybody, but I see, you know, people, the legal documents that sometimes are sent to me, it's just not there yet. Whether it gets there in the next few years, I don't know, but I just can't see. The human element is just so important because it looks backwards rather than forward. Right. And so great minds think forward, not backwards. And AI is incapable of thinking forwards.

Speaker A: Well, that is a great place to end our conversation. We're all going to be thinking forward and continuing to lean into what makes us uniquely human. Candace, thanks so much for being our guest today. If you of, uh, where you're speaking next, you want to throw out a couple of places.

Speaker B: Uh, yes, I will be at the ANA conference in Huntington beach next week and then I will be at the Ad Age Small Agency Conference in July. And uh, I think I have a webinar that's also coming up, but those are the two that are probably, uh, currently being advertised.

Speaker A: Awesome. And if you're interested in hearing more from Candice, you can find her on LinkedIn or on her firm's website. She has a fantastic bio page and we'll, uh, have the full transcripts on modop.com you can also find more episodes there or wherever you listen to your podcast you can Search Leader Generation until next time, Candice. We'll have to see what happens in another couple of years. Thanks again for being our guest.

Speaker B: We'll be back. Thank you, Tessa. Take care.

Speaker A: You too.

Speaker B: Bye.

Speaker D: A Leader Generation podcast is brought to you by ModUp, a full service marketing communications agency focused on using the right methods to help clients capitalize on their opportunities. Modop services for both B2C and B2B markets include brand strategy, advertising, digital marketing, lead generation and public relations. Through its technology group, modop also offers website e commerce and app UX design, development and technology stack integration. For more information, visit modop.com that's m o d o p dot com.