AI Copyright Law

Episode 59: AI Copyright Law, with Sharon Toerek

podcast photo thumbnail



View this episode on:

AI Copyright Law? This is your comprehensive guide to the legal rights surrounding Artificial Intelligence. Stay informed about AI Copyright Law.

AI copyright law—I am so excited to welcome Sharon Toerek back to the Sell With Authority podcast. If you’re meeting Sharon for the first time, she’s the Principal of Toerek Law and the host of the rock-solid awesome podcast The Innovative Agency.

And in full transparency — Sharon is Predictive’s wicked, smart intellectual property attorney. Over the years — she has brilliantly helped us protect our intellectual property so that as it evolved and morphed into revenue streams — we were ready.

In Episode 52 of Sell With Authority, Sharon broke down how to protect your intellectual property and turn it into revenue streams.

Today, Sharon is back for an encore episode where we will explore the legal implications of using AI to create content for client projects. We discuss copyright, trademark, and other intellectual property rights and consider several questions that arise when using AI-generated content.

Sharon brings her expertise to discuss AI copyright law in the workplace. Her advice will help agencies navigate the challenges surrounding AI and protect their intellectual property.


What you will learn in this episode is about AI copyright law:

  • Are there legal implications to using AI to create content as part of a client project
  • Who owns the content generated by AI
  • Questions agencies should consider regarding AI copyright law, trademark, and IP rights
  • What happens when AI-generated content is modified by an agency as it relates to intellectual property rights
  • How can clients be assured that the content, campaigns, or brand they have as a result of working with an agency is truly theirs


Additional Free Resources:

  • Subscribe to our new “Sell with Authority” podcast. Learn how to attract a steady stream of well-prepared prospects into your sales pipeline and convert them into right-fit clients. Everything you need to move further faster is shared in full transparency. Go here for all the episodes.
  • Order your free copy of “Sell with Authority” in paperback or Kindle. Honest — no hoops to jump through. All you need to do is go here and enter your mailing address. No credit card for shipping. Nothing.
  • Join our free “How to Fill Your Sales Pipeline” Facebook Group. Over 1,300 agency owners just like you are looking to move further faster by sharing best practices and what’s working for them. But — as Susan Baier recently said, “It’s not about the numbers, it’s about the SMARTS contributed to this wonderful, generous community!” Go here for details.


AI Copyright Law: Full Episode Transcript


Welcome to the Sell with Authority podcast. I’m Stephen Woessner, CEO of Predictive ROI, and my team and I created this podcast specifically for you. So if you’re an agency owner, a business coach, or a strategic consultant, and you’re looking to fill your sales pipeline with a steady stream of right-fit prospects, you know, to get the at-bats that you need in order to build and scale, then you’re in the right place. Do you want proven strategies for becoming the known expert in your niche and attracting all the clients you need? Yep. We’re gonna cover that. You wanna learn how to step away from the sea of sameness so you can actually stand out and own the ground that you’re standing on. Yep. We’re gonna cover that too. Do you want to future-proof your business so you can successfully navigate the next challenge that comes your way?


Well, absolutely. We’ll help you there as well. I promise you each episode of this podcast will contain valuable insights, tangible examples, and best practices, never theory, from thought leaders, experts, and owners who have done exactly what you’re working hard to do. So, I want you to think practical and tactical. Never any fluff. Each of our guests has built a position of authority and then monetized that position by growing their audience, nurturing leads, and, yes, converting sales. But all the while, they did it by being helpful. So every time someone from their audience turned around there, they were with a helpful answer to an important question. So the RightFit prospects never, ever felt like they were a prospect. I also promise you that every strategy that we discuss and every tool we recommend will be shared completely transparently in each episode so that you can become a known expert in your niche.


Learn more about AI Copyright Law by listening to our first interview with Sharon Toerek


AI Copyright Law: Sharon’s Introduction


So you can fill your sales pipeline with a steady stream of RightFit clients and then never ever make them feel like they’re one of your prospects. Okay, so I am super excited for you to meet our very special guest expert today, Sharon Toerek. So, if you’re meeting Sharon for the first time, she’s the principal of Toerek Law and the host of the Rock Solid awesome podcast, the Innovative Agency. In full transparency, Sharon is a predictive, wicked, smart intellectual property attorney. So, over the years, she has brilliantly helped us protect our IP so that as it evolved and morphed into revenue streams, we were always buttoned up and ready. So, if you haven’t yet listened to episode 52 of the podcast, I would highly recommend it because Sharon broke down how to protect your intellectual property, as well as generously, and then provided a helpful framework for how to turn your IP into revenue streams.


So, we will include a link to episode 52 of the podcast in today’s show notes. But I invited Sharon back for today’s encore to continue our intellectual property discussion but with a twist. More specifically, Sharon and I are going to wade through the mist, maybe the swamp, the myriad of questions surrounding AI and intellectual property rights. For example, are there any legal implications to using AI to create content as part of one of your client projects? And if so, what questions should we as agencies be considering or asking ourselves as it relates to copyright, trademark, and other intellectual property rights? For example, who owns the AI-generated content? Does ownership belong to the company that owns the AI tech? Wouldn’t that be a kicker? Does the agency that created the successful AI prompt, does it own the content?


Here’s more information about AI Copyright Law by Sharon Toerek


AI Copyright Law: Navigating Ownership and Rights


Or is it the client who engaged the agency to create the work? Are they the rightful owner? Or could it be nobody since there was nobody, as in a person who created it? And what if the content that was generated by AI by your agency team was then modified? Maybe you cleaned it up, or maybe you improved it. Does that make a difference as it relates to intellectual property rights? Or how about this curve ball? What if your agency used AI as an inspiration for a design, and then your designers created the actual design, but the finished work looks awfully close to the AI generator version? So who owns it? Or how can clients be assured that the content, the campaigns, the brand, and the litany of other things that they have as a result of working with an agency are indeed owned by them once the bills are paid and the agency hands over all of the assets?


Again, Sharon and I are gonna wade through this muck and mire, and I have no doubt that several of these, uh, questions are gonna represent very sticky wickets, which is why having Sharon’s expertise to lean on is going to be so valuable. The last thing you want to do in this exciting time of AI, and it is indeed exciting, is to unintentionally put your agency in hot water with a client. I promise you that applying what sharing gives you during this episode will help you protect your intellectual property and protect your agency at the same time. Okay. So, without further ado, welcome back to the Sell with Authority podcast, Sharon.


Woo. No pressure. Thank you, .


We just have to solve all the world’s problems here, Sharon. I don’t wanna, we’re gonna have to get a ticker counting. Um, the times I use the word depend, probably in the answering of the questions that you’re gonna be bing at me. I am, I’m super delighted to be back talking with you, my friend, about this. It is indeed an exciting time. Um, and there’s, there are lots of things to think about, about how this is gonna affect agency life and work delivery, and it’s, it’s happening more quickly than I think any of us predicted. And so that, that makes it, um, super challenging as well.


Be updated on AI Copyright Law by reading this blog from Sharon Toerek


AI Copyright Law: AI-Generated Content on Intellectual Property Rights


Amen to that. Uh, which, again, I’m super excited that we get to have this conversation, uh, ’cause I know your advice is gonna be helpful, and yes, I totally get that. There’s going to be a few depends on this, or it depends on that lot, lots of, lots of depend in there, which I certainly understand. Let, let’s go, let’s go high level first because I know you’re looking at this closely. Let, let’s go to the high level first, and then we’ll go eye level into some more of the muck and the Myers as I framed it up. But, let’s go high level. So what are some of the higher level implications that you see as it relates to AI generated content and, uh, intellectual property rights?


So I think the first thing that concerns me for agencies from an IP dash legal perspective is the expectations that you set with your client, um, when you’re planning the work, to begin with. And, perhaps, as early as when you are negotiating the terms of your agreement with the client or the scope of work that you’re gonna be doing for them. The reason I say that is because I think a lot of the tension between agency and client can be reduced around this whole question of ownership of AI-generated content with, first of all, transparency, which should be your hallmark anyway, but then also some vulnerability and, um, some reflection of that transparency and vulnerability in the, in the documents that you sign and the written commitment commitments that you make to the client. And let me, I guess, explain a little bit what I need.


We’re in a state right now where it’s very clear that machine-generated content can’t be owned by anybody. The copyright offices, to the extent they’ve been clear about anything, they’ve been clear about the fact that if it came from a machine, it was generated from a machine, there’s no copyright stake in that, in that output. Oh, that sounds simple. But when you layer on all these other contingencies, what about your prompt strategy? What about the extent to which you edit or vary the content once it’s generated, um, from AI? What about the human insight that goes into deciding how to use all of that content? Do all those add additional layers of complexity? And so I think, you know, one of the first high-level things to think about is, is what are you going to say, represent, and promise to your client at the beginning when you are planning and sort of iterating your strategy about how you intend to use AI generative tools?


Learn more about AI Copyright Law by listening to our first interview with Sharon Toerek


AI Copyright Law: Documenting AI Usage


AI copyright law? What proportion of the work product do you expect to be the output of AI generate, uh, of ai, mm-hmm, and to what extent will then that output be, um, adjusted, edited, reflected by human intervention? And are you going to use it for any underlying research or fact gathering that will end up reflected in the final work product where, you know, truthfulness, veracity, and accuracy are gonna be key because more than being worried about ownership of IP and the final product, your clients certainly don’t wanna be taking on any more risk reliability out in the consumer marketplace or in their customer marketplace as a result of any marketing activity you’ve done for them. Um, whether it’s advertising, creative, whatever it might be. And so your contract needs to be transparent on these points. We plan to use these tools; we expect that approximately x percentage of the output, um, will be generated for, and this is how we’re gonna use them. These are the tasks we’re gonna use them to do. This is how we’re, so that’s where you start, I think, at a high level. I guess I’ll just let that settle there for, for reaction, Stephen, but I think we’re so worried about the risk that we’re taking and maybe these big questions around ownership that we might be missing a critical misstep in the agency-client relationship, which is that conversation about this huge vast unknown regarding, um, is this really gonna be faster, cheaper, better than the entirely human-generated output we’re used to?


AI copyright law? So this I was hearing you say that I wrote in my notes here, like when you were mentioning the contract and you mentioned such keywords like, uh, transparency and vulnerability. I love that. And then you mentioned the contract. So I was thinking like, you know, just like an MSA master services agreement might say, you know, X percentage of the work we anticipate are going to be, is going to be performed by subcontractors. Right? Like, like typically, agencies have to disclose that and, and then in, sometimes in some instances, right? Uh, an agency has to get a client’s permission Yeah. In order to outsource something as part of the scope of work. And but it has to be documented in the MSA, is that what you’re suggesting here? That that AI also needs to be documented like that, just like we would for a subcontractor?


Yeah, I think that’s, that’s a perfectly APTT analogy. I mean, what we’re talking about is third-party-generated content, right? In this case, it just happens to be a machine that is creating it. But just as your master service agreements should allow for an acknowledgment by the client that you may use third-party sources or third parties themselves, freelancers, contractors, and then think about all other, all the other third-party sources you are already using as an agency to prepare the deliverables, stock photography, um, software code, um, stock video, font, graphics. These are things that the agency didn’t create and doesn’t own the rights to. You’re licensing them, or you’ve engaged a third party, um, to help create them for you to, um, integrate them into the deliverable. Mm-Hmm. Your contract also needs to include the variable in that section of AI-generated content so that there’s a place to trigger the conversation between the agency and the client.


Here’s more information about AI Copyright Law by Sharon Toerek


AI Copyright Law: Ownership and Liability in AI-Generated Content


What about AI copyright law? How and what tools do you plan to use? How much of the final output do you think will be AI-generated? And those conversations aren’t necessarily gonna all be reflected verbatim in the contract or the SOW, but the acknowledgment and the agreement that the agency will use third-party resources, um, such as generative AI to create some of the underlying content or some of the final deliverables, whatever the case might be. So that there’s no doubt about the fact that the parties are both assuming some risk here, um, with respect to the ownership question and with respect, with respect to the liability question related to, um, you know, claims facts cited, um, and any inadvertent infringement with other stuff that’s out there in the marketplace.


Yeah. And, and so, okay, so that, that, let’s go into the ownership piece, and then that maybe takes us back into the infringement piece. So if there’s no ownership, if I’m understanding you correctly, and if there’s no ownership over the AI-generated content if the U-S-P-T-O has been, has been clear about that, then who, who, who does own it? And how do we protect ourselves? I mean, it’s a prompt and a query, right? Then, the machine or AI-generated content is spat out. So, how are we sure that it isn’t going to infringe on somebody else’s prompts that generated the same thing? Like how do we know?


Right. There are a couple of questions in your question, really. Okay. Um, I think what the copyright office has been clear about so far, and it’s a very narrow lane of clarity, is that content that is entirely AI-generated, the output there, um, is not protectable under copyright law. For example, if you prompted it to write, um, a piece of ad copy or you prompted it to write a white paper about a subject, that output would not be, you can’t slap your copyright notice on it and claim any sort of ownership because it was not generated by a human. Okay. That’s a very pure and narrow lane of example. Correct. Okay. But what no one is addressing yet, and there’s no case law yet to address it, I expect that we’ll develop over time. And I expect that as regulatory oversight catches up with this, which is, I think gonna take a while ’cause it is moving so fast that, you know, I have always said technology goes first, business sort of runs to keep up and then the law, um, sweeps up the mess left behind.


Be updated on AI Copyright Law by reading this blog from Sharon Toerek


AI Copyright Law: Protecting Prompt Strategies


AI copyright law? And that’s the order in which it happens. And since the technology itself is, is iterating so quickly, even the, even business and industry are having trouble, you know, keeping up with using it intelligently and smartly and to any significant effect. And then, you know, here, here are US lawyers with our running shoes on trying to keep up with you all and, um, figure what all this means. So let’s talk about, um, some adjacent issues from an IP perspective. Let’s talk about the prompt strategy. Hmm. Right now, as we’re sitting here having this conversation, I firmly believe that the extent to which your prompt strategy is unique is uniquely informed by the skill you have as an agency either to serve a particular vertical industry or in a particular discipline of marketing, whatever that might be. All that informs your prompt strategies, your prompt strategies can and should be original.


And so are those protectable. And what I maintain right now is that those feel to me like any other trade secret could, uh, feels and could be protected. Wow. So the degree to which though, you can protect it depends upon the degree to which you maintain it as a secret, right? Because trade secrets have to be beneficial and valuable because of their secret status. So if you are whizzbang at developing great prompt strategies that create sound output, those, I think ultimately, will end up being as valuable as some of the output is depending on what we’re creating with the generative AI. So, can you own the rights to those prompt strategies? I don’t see any reason why you can’t, again, as long as you follow the principles of trade secret maintenance, and trade secrets is a, is a form of intellectual property that’s probably least discussed in the marketing world and the agency world because agencies give away ideas so freely and discuss ideas so freely. But if you, if you want to own your prompt strategies and if there’s something unique and special sauce about them, I see no reason why, again, there’s no case law on this yet, but they, it feels very much to me like any other strategic advantage or competitive advantage you have because of something you know that your competitor does not know. So, prompt strategy.


Oh, okay. So hang, hang on just one second. So let’s, let’s loop back to episode 52. Mm-Hmm. When you were sharing your smarts around IP to income, right? So let’s say that, let’s say that predictive ROI came up, you know, ’cause we love to run stuff through our lab, and one of the lab experiments was prompt strategies, and we came up with, you know, whatever topic, that kind of stuff. A series of prompt strategies Mm-Hmm. So, we could protect that, and maybe we could then sell that as a product like we talked about in episode 52. Am I tracking with you or not like that?


Learn more about AI Copyright Law by listening to our first interview with Sharon Toerek


AI Copyright Law: Ownership and AI-Generated Output


AI copyright law? I think kind of, I mean, ’cause we talked about, I mean, one of the, I don’t, I don’t, I won’t rehash it ’cause everybody should go back and listen, but one of the, uh, parts of the framework for assessing your IP for monetization is the content piece. Typically, you monetize content, uh, or you protect, I should say content either under copyright law, which the copyright office is already giving us guidance that the output that is all machine-generated isn’t protectable as under copyright law. Um, or in rare circumstances, your content. And in this case, the content I refer to wouldn’t be the output, but it would be the prompting strategy Hmm. To be protected as a trade secret. Wow. Wow. And the way you do that is under confidentiality agreements or simply by maintaining it and not disclosing it outside the four walls of the agency.


Admittedly, from a practical perspective, I’m not gonna lie, it’s hard to do. Um, but we’re talking theoretically and what the possibilities are, suitable? The universe of IP potential protection for mm-hmm. , um, stuff that an agency generates in a typical day or could with, um, with ai. Hmm. So think about that. That’s, that’s prompting. So that’s at the inception phase of creating some deliverables using generative AI. So what about the output itself? Well, we already kind of talked about the pure output and what the copyright office says about that, but what about the degree to which humans who work for your agency are changing? Hmm. Editing it, um, creates an entirely new work based upon the inspiration that they get from the output. So there’s gonna depend upon the degree of, um, human creation of that end product. And, so if you’re simply using it for inspiration and you create an entirely new work that is not duplicative simply of what you’ve, um, gotten from generative ai, conceivably you could own the copyright in that work.


Hmm. Um, I say conceivably because I don’t wanna say depend and because there’s no case law yet on it, or, and there’s no regulation or statute on this, but the, in this, in that case, you’ve got evidence of a lot of human hands on creating the content, and you use the AI output as inspiration. Just might you, like you might use a piece of art, or you might use an event you attended, or a piece of music where it gets slippery, um, is where you can’t really draw a clean line between the output from AI and what the humans have contributed to the final work product. So I think one, you need to decide how much this matters to you if you’re an agency using AI to create storyboards, emails, or blog posts. I think a lot of that work’s gonna dry up anyway because I think people are under the mistaken impression that AI tools are gonna replace human involvement in that.


Here’s more information about AI Copyright Law by Sharon Toerek


AI Copyright Law: Navigating the Legal Landscape


I’m not convinced because I think a lot of the content that’s being generated right now is not quality, and it’s gonna be ubiquitous and not distinguishable. So, for the time being, I mean, that will change, but for the time being, I think you still need human insight into the final deliverables. And so what proportion of human insight and what proportion of human involvement will matter? I can’t give you a percentage, uh, point, but I think that will matter. Um, in terms of determining ownership status. And then infringement, you mentioned the impossibility, uh, I think in advance about reducing the risk of infringement liability, the laws regarding AI copyright law have not changed because of the emergence of generative AI. Hmm. If you do, there are just two elements to proving copyright infringement. You have access to an underlying work, and the work that you create is substantially similar to that original underlying work.


So, is it possible for you to end up independently, let’s forget AI for a second. Could you create a piece of content, whether it is a logo or a blog post, that looks, um, startlingly similar, substantially similar to somebody else’s underlying work, and not be guilty of copyright infringement? Yeah, you could, if you never saw the underlying work. But if you saw it, if you had access, and your work is substantially similar, then there’s potential copyright infringement liability. How will the two of you connect with each other so that someone knows they’ve been infringed upon? That’s an open question. Hmm. Right. So, I think you have to decide what level of risk you’re comfortable with. Okay. I think you have to decide whether it ultimately matters to, and I say that I use this example a lot in the agencies we counsel, um, who do influencer marketing.


There’s a lot of battle back and forth between influencers, their agencies, and the brand and the brand’s agency who’s gonna own the content. And I can’t tell you how many times I’ve said to them, all this is a Christmas campaign. By the time February 1st comes around, who is gonna care who owns the rights to this underlying content? Right. Really? Um, ’cause most of the time, it’s just negotiating leverage, um, in a conversation. Similarly, does it matter to you as a brand, as long as you’re not infringing or you’re indemnified against infringing, if you own the IP and the output that generative AI creates, if it does matter to you, okay, then we need to take some additional steps perhaps to reduce risk. If it doesn’t matter, perhaps your risk isn’t that high. I think, you know, you really have to constantly be thinking about how much risk are we really assuming and what’s our tolerance for risk? ’cause, as I said, so there’s still a lot of open questions.


Be updated on AI Copyright Law by reading this blog from Sharon Toerek


AI Copyright Law: Human Collaboration and AI Inspiration


Well, this is so good. Let’s go back to the AI copyright law piece when you mentioned depends on the human creation, the, uh, contribution. Let’s assume that the AI generative content serves as the inspiration, and then we stack on top of it the human collaboration and make it a better kind of stuff or whatever. Right? Then, we put our AI copyright law on it. But in defending that copyright, my assumption is, and that, that we would then, like part of that defense would, we would have to disclose that 30% of it, I’m sure I’m not saying this right, like 30% or 40% of the original work, or excuse me, the finished work that inspiration came from ai. Like, is that part of then the defense and the ownership of the copyright?


I think so. I think, you know, the way that a copyright office traditionally has distinguished is not so much in percentages, but if, for example, a, a great comparison here would be a book where one person has written the copy and another has illustrated with original drawings. Okay. So the, uh, another great example is, um, musicians who own the publishing rights to their work. So, they own the compositions themselves versus the sound recording of them performing the work. Those are two separate. So, I think with a copy, I think it is possible for you as an agency to create a piece of work that is partially inclusive of AI-generated work and then add to your original contributions to that work. So, you may have used AI to generate a script, but you’ve contributed the original imagery. Okay. Or vice versa. No one is saying so far that you can’t own the rights to your human contribution to the final work. You would just have to disclaim anything that the AI copyright law. So, for example, if I am pursuing a copyright registration in, um, a white paper, for example, and its original research, but I fight, I might have fed the research results into an AI generation tool, and the executive summary that it spits out might be okay with a bit of polishing. So what I’m, what might I do, I might, um, file for copyright of the report as a whole and exclude the rights to the executive summary. Ah,


Okay. Or disclaim any images if they were AI-generated. So far, the copyright office is saying that if you can distinguish between the two and disclaim that which is not human-generated, you can still own the copyright in its human-generated components. So you do have options if the dividing lines are clear, it’s when you blur a lot, and then you make it an alphabet soup that it’s gonna be very hard to distinguish between what humans created, what the AI-generated, and therefore, who owns the final work product. Um, that will be a more challenging position to take. Mm-Hmm. Either way.


Learn more about AI Copyright Law by listening to our first interview with Sharon Toerek


AI Copyright Law: AI in Academic Research


A few weeks ago, I saw some examples of, or I should say, not examples, like use cases of chat GPT, and one of those was in the academic realm. I know that’s tangentially kind of connected to what we’re talking about here, but it’s about thought leadership. And so, uh, in the academic realm where like if I’m writing a scholarly, if I’m writing an article or research paper for a scholarly journal, peer-reviewed scholarly journal, one of, one of the things that takes the most time, as you know, is the literature review and then writing the abstract and getting that really detailed so that then, you know, my colleagues and I can add some stuff on top of that. And now it’s a new work that we could get submitted for publication. And so the use case was, yeah, instead of doing that, take the 10 best papers that would be part of your literature reviews, throw ’em into chat, GPT, ask chat GPT to create the abstract, and then you put your stuff on top of it. And then, as I was thinking about that, now hearing you share this with us and your insights, I’m thinking, well then, wouldn’t that have to be disclosed to the journal that you’re submitting to for publication? That part of the work, maybe 20, 30% of the literature review, was actually done by chat GPT.


AI copyright law? Yeah. I think, you know, ethically definitely, and legally just for the purposes of, uh, risk reduction for yourself as a, uh, an agency disclosure, that vulnerability and disclosure we talked about, you know, moments back and making sure that your contract with your client reflects the fact that, and you may not know the exact percentages or you may wanna cite it as a range. I’d put that in the SOW, frankly, not, not, I wouldn’t put the number itself in the master service agreement, but that’s a choice that I would make. You know, yes. You, you should, uh, make those disclosures to, and, and if the, if that balance changes once you get into the project, then you should have that discussion with the client and reflect that discussion in a change order or in a new SOW so that, um, you’re, you’re not put in a position as an agency where you deliver the content and someone calls the client on the carpet for it after it’s released out into the wild, and they come to you and say, what about this? There’s, this isn’t, and well, yeah, we told you that we were using AI generation tools and, um, this is sometimes what happens and, and we don’t know how frequently it’s gonna happen. We can only guess it’s gonna increase in frequency as AI gets smarter.


So, I would love to get your perspective on this because you walk alongside agencies every single day talking about sticky wickets like this. Yeah. So if you were to just come up with a percentage, like what percentage of agencies do you think are actively thinking about this and, and coming to you as saying, Hey, Sharon, can you help us update our MSA? Because you know, we, we think that we’re exposed here. How many agencies do you think are up to speed on this? If, if like, best guess?


Here’s more information about AI Copyright Law by Sharon Toerek


AI Copyright Law: Addressing AI Ownership in Agency Contracts


AI copyright law? I think the number’s increasing. I still wouldn’t put, I wouldn’t peg it yet at half, um, based on the communications we’re having so far. But, um, but it could get close to that quickly. I think agencies are sort of curious, but not necessarily. But I think they’re more curious about it from an ownership perspective, not necessarily curious about what it means in terms of what they should be saying in their legal documents about it. Okay. Um, but we are getting a rising number of requests from clients, even agencies for whom we’ve created their MSAs in the past, saying, Hey, I think we ought to revisit this and address AI specifically in the agreement. And so, you know, we are working on a way to do that that can be universally applied so that we can say, Hey, include this language in the section of your contract where you talk about third party works, ah, which is I think right now where it belongs.


AI copyright law is essential to know because AI is a third party, not a human third party, but neither is a stock photography service or software, you know, a third-party software. So I think that this, that’s where this goes, and I think that that’s where this discussion goes. And I just think it needs to be called out more than an agency. Agencies don’t love. And this has, you know, been part of my soapbox for the years that I’ve been counseling them. Agencies don’t love spending a lot of time thinking about IP ownership regarding the work that they do, mainly if they’re doing it, as work that is customized to be owned by the client. At the end of the day, um, agencies do think about creating their own IP, sometimes, not as much as I’d like them to, but they do. And how they would monetize that, turning it into revenue streams and such.


But they don’t love thinking about because if, if in their mind, at the end of the day, the client’s gonna own it, um, they’re not thinking a lot about the, the protection rights. They indeed don’t wanna get sued for plagiarism or infringement. But, um, beyond that, they’re, you know, traditionally not so worried this has gotten them thinking about it. Um, and so it’s an opportunity to look at your whole IP strategy at a high level, um, and making sure that your agreements not only with your clients but with the independent contractors and freelancers who you engage mm-hmm, um, you need to know and ask questions about as an agency, what are you doing with AI to generate the work that you deliver to us that we then incorporate into the client deliverables? Um, because we need to know, we need to be in a position to disclose, and we need your representations about what you’re doing so that everybody is clear on that point. And, you know, you also need to add another domino and the potential chain of liability for infringement or using unauthorized work.


Be updated on AI Copyright Law by reading this blog from Sharon Toerek


AI Copyright Law: Key Considerations for AI in Agency Work


This is so brilliant, uh, as, as always. And I cannot wait to share with, uh, Hannah on our team who is, you know, running our predictive lab about the prompt strategies, and I’m like, holy bananas. Do we have an idea of an experiment to run through the lab? Uh, that was really, really cool. So I, I know we need to come in for, uh, a landing here, but before we go, before we close out and say goodbye, um, Sharon, any, any final pieces of advice, anything you think we might have missed? Any sticky wickets that are still out there that we should talk about? And then please do tell our audience the best way to connect with you.


I appreciate it, and thanks again for inviting me. I always love having conversations with you, Stephen. Um, they, uh, they get me, they make me more intelligent too. So I would like to have the opportunity to talk about this stuff. So I think number one, the things you should expect to have to deal with in your client conversations, what are you gonna represent about the number, the percentage or amount of deliverables that could be AI-generated? Um, what are you gonna prompt your client to think about in terms of ownership of the output of anything that’s AI-generated? Um, number two, I think with respect to IP ownership, be thinking proportionately about, um, to the extent you can, how much you might be relying on AI to generate the output. Be thinking about your workflow and that relates to the flow, the prompt strategy discussion that we had.


AI copyright law, at what stages might you use it, what are you gonna be putting into it, how original is that and how will you protect it? And then managing the legal implications of the work once it gets out into the world. Be thinking about adding to your workflow places for verification of the accuracy of the content that is turned over to the client to put out into the world and be reminded. Be mindful of the fact that using AI to do your background research or together inspiration is not, um, a substitute for due diligence. You still need to be doing your trademark searching. You still need to be thoughtful and have integrity regarding the originality of the human-created content that you add to the AI. Um, and to the extent you know about substantial similarity, even if you’ve used AI, you need to think about the implications of infringement there.


Learn more about AI Copyright Law by listening to our first interview with Sharon Toerek


AI Copyright Law: Legal Insights and Best Practices


So I would say that’s a pretty extensive laundry list and a high level of things to think about as an agency. And I think those are your starting points and that stuff that agencies at, at any size and scale, I think and devote some time and energy around. And we can be [email protected]. Um, and I also spend quite a bit of time on LinkedIn. Um, and we do, uh, in full transparency with help from our good friends at predictive on quarterly q and sessions around legal, uh, issues. We have focused them on the essential legal documents agencies need. But we will be weaving AI conversation ’cause we don’t have a choice but not to, um, it’s in, it’s crucial to the work you’re all doing. Um, and it’s exciting, and so, um, it’s fun to be on top of. So keep your eyes out for those by either, um, connecting with me on LinkedIn or reaching out via email and we’ll get you on our list so that you get an invite when we have the next one. I’ll be at Baba to build a better agency summit. I know Steve, me too excited about that. So connect with me in person there if you’re gonna be there.


Well, okay, so that’s a great point and, and reminder because if you, if you happen to be listening in to this the day that it, uh, aired, which is, so today would be May 17th, if you’re listening to this now, uh, Sharon is at Build a Better Agency Summit. Uh, so please be sure to look her up. ,


I will be there, and I will be actually leading round tables on artificial intelligence for agencies—three of them, I think. So yeah, come say hi.


Amazing. Okay, everyone, no matter how many notes you took or how often you go back and re-listen to the words of wisdom that Sharon just shared with you. The key is you have to take everything that she just shared and apply it. You have to take it and apply it. And if you do, not only will you accelerate your results, but you’ll have a much more protected agency. So, Sharon, we all have the same 86,400 seconds in a day. And I am grateful that, again, you said yes to come onto the show, to be our generous guide and mentor to help us move our businesses onward to that next level. Thank you so much, my friend.


Absolutely. It is my pleasure, Stephen. It is always a delight to talk with you. Thank you for having me again.

Fill Your Sales Pipeline Q&A

LIVE Wednesdays at 1:00 pm Eastern / 12 Noon Central

Sell with Authority Podcast

The Sell with Authority Podcast is for agency owners, business coaches, and strategic consultants who are looking to grow a thriving, profitable business that can weather the constant change that seems to be our world’s reality.

Follow Us