In-Ear Insights: AI And the Future of Intellectual Property

In this episode of In-Ear Insights, the Trust Insights podcast, Katie and Chris discuss the present and future of intellectual property in the age of AI.

You will understand why the content AI generates is legally unprotectable, preventing potential business losses. You will discover who is truly liable for copyright infringement when you publish AI-assisted content, shifting your risk management strategy. You will learn precise actions and methods you must implement to protect your valuable frameworks and creations from theft. You will gain crucial insight into performing necessary due diligence steps to avoid costly lawsuits before publishing any AI-derived work. Watch now to safeguard your brand and stay ahead of evolving legal risks!

Watch the video here:

Watch this video on YouTube

Can’t see anything? Watch it on YouTube here.

Listen to the audio here:

Download the MP3 audio here.

[podcastsponsor]

Machine-Generated Transcript

What follows is an AI-generated transcript. The transcript may contain errors and is not a substitute for listening to the episode.

Christopher S. Penn: In this week’s In Ear Insights, let’s talk about the present and future of intellectual property in the age of AI.

Now, before we get started with this week’s episode, we have to put up the obligatory disclaimer: we are not lawyers. This is not legal advice. Please consult with a qualified legal expert practitioner for advice specific to your situation in your jurisdiction.

And you will see this banner frequently because though we are knowledgeable about data and AI, we are not lawyers.

We can, if you’d like, join our Slack group at Trust Insights, AI Analytics for Marketers, and we can recommend some people who are lawyers and can provide advice depending on your jurisdiction.

So, Katie, this is a topic that you came across very recently. What’s the gist of it?

Katie Robbert: So the backstory is I was sitting on a panel with an internal team and one of the audience members.

We were talking about generative AI as a whole and what it means for the industry, where we are now, so on, so forth.

And someone asked the question of intellectual property. Specifically, how has intellectual property management changed due to AI?

And I thought that was a great question because I think that first and foremost, intellectual property is something that perhaps isn’t well understood in terms of how it works. And then I think that there’s we were talking about the notion of AI slop, but how do you get there?

Aeo, geo, all your favorite terms. But basically the question is around: if we really break it down, how do I protect the things that I’m creating, but also let people know that it’s available? And that’s.

I know this is going to come as a shocker. New tech doesn’t solve old problems, it just highlights it.

So if you’re not protecting your assets, if you’re not filing for your copyrights and your trademarks and making sure that what is actually contained within your ecosystem of intellectual property, then you have no leg to stand on. And so just putting it out there in the world doesn’t mean that you own it.

There are more regulated systems. They cost money. Again, as Chris mentioned, we’re not lawyers. This is not legal advice. Consult a qualified expert.

My advice as a quasi creator is to consult with a legal team to ask them the questions of—let’s say, for example—I really want people to know what the 5P framework is. And the answer, I really do want that, but I don’t want to get ripped off. I don’t want people to create derivatives of it. I don’t want people to say, “Hey, that’s a really great idea, let me create my own version based on the hard work you’ve done,” and then make money off of you where you could be making money from the thing that you created.

That’s the basic idea of this intellectual property.

So the question that comes up is if I’m creating something that I want to own and I want to protect, but I also want large language models to serve it up as a result, or a search engine to serve it up as a result, how do I protect myself?

Chris, I’m sure this is something that as a creator you’ve given a lot of thought to. So how has intellectual property changed due to AI?

Christopher S. Penn: Here’s the good and bad news. The law in many places has not changed. The law is pretty firm, and while organizations like the U.S. Copyright Office have issued guidance, the actual laws have not changed. So let’s delineate five different kinds of mechanisms for this.

There are copyrights which protect a tangible expression of work. So when you write a blog post, a copyright would protect that. There are patents. Patents protect an idea. Copyrights do not protect ideas. Patents do. Patents protect—like, hey, here is the patent for a toilet paper holder. Which by the way, fun fact, the roll is always over in the patent, which is the correct way to put toilet paper on.

And then there are registrations. So there’s trademark, registered mark, and service mark. And these protect things like logos and stuff, brand names. So the 5Ps, for example, could be a service mark. And again, contact your lawyer for which things you need to do.

But for example, with Trust Insights, the Trust Insights logo is something that is a registered mark, and the 5Ps are a service mark. Both are also protected by copyright, but they are different. And the reason they’re different is because you would press different kinds of lawsuits depending on it.

Now this is also, we’re speaking from the USA. Every country’s laws about copyright are different.

Now a lot of countries have signed on to this thing called the Berne Convention (B E R N, I think named after Switzerland), which basically tries to make common things like copyright, trademark, etc., but it’s still not universal. And there are many countries where those definitions are wildly different.

In the USA under copyright, it was the 1978 Copyright Act, which essentially says the moment you create something, it is copyrighted. You would file for a copyright to have additional documentation, like irrefutable proof. This is the thing I worked on with my lawyers to prove that I actually made this thing. But under US law right now, the moment you, the human, create something, it is copyrighted. Now as this applies to AI, this is where things get messy.

Because if you prompt Gemini or ChatGPT, “Write me a blog post about B2B marketing,” your prompt is copyrightable; the output is not.

It was a case in 2018, *Naruto vs. Slater*, where a chimpanzee took a selfie, and there was a whole lawsuit that went on with People for the Ethical Treatment of Animals. They used the image, and it went to court, and the Supreme Court eventually ruled the chimp did the work. It held the camera, it did the work even though it was the photographer’s equipment, and therefore the chimp would own the copyright. Except chimps can’t own copyright.

And so they established in that court case only humans can have copyright in the USA. Which means that if you prompt ChatGPT to write you a blog post, ChatGPT did the work, you did not. And therefore that blog post is not copyrightable.

So the part of your question about what’s the future of intellectual property is if you are using AI to make something net new, it’s not copyrightable. You have no claim to intellectual property for that.

Katie Robbert: So I want to go back to I think you said the 1978 reference, and I hear you when you say if you create something and put it out there, you own the copyright. I don’t think people care unless there is some kind of mark on it—the different kinds of copyright, trademark, whatever’s appropriate. I don’t think people care because it’s easy to fudge the data. And by that I mean I’m going to say, I saw this really great idea that Chris Penn put out there, and I wish I had thought of it first. So I’m going to put it out there, but I’m going to back date my blog post to one day before.

And sure there are audit trails, and you can get into the technical, but at a high level it’s very easy for people to say, “No, I had that idea first,” or, “Yeah, Chris and I had a conversation that wasn’t recorded, but I totally gave him that idea. And he used it, and now he’s calling copyright. But it’s my idea.” I feel unless—and again, I’m going to put this up here because this is important: We’re not lawyers. This is not legal advice—unless you have some kind of piece of paper to back up your claim. Personally, this is one person’s opinion. I feel like it’s going to be harder for you to prove ownership of the thing. So, Chris, you and I have debated this.

Why are we paying the legal team to file for these copyrights when we’ve already put it out there? Therefore, we own it. And my stance is we don’t own it enough.

Christopher S. Penn: Yes. And fundamentally—Cary Gorgon said this not too long ago—”Write it or you’ll regret it.” Basically, if it isn’t written down, it never happens. So the foundation of all law, but especially copyright law, is receipts. You got to have receipts. And filing a formal copyright with the Copyright Office is about the strongest receipt you can have. You can say, my lawyer timestamped this, filed this, and this is admissible in a court of law as evidence and has been registered with a third party.

Anything where there is a tangible record that you can prove. And to your point, some systems can be fudged. For example, one system that is oddly relatively immutable is things like Twitter, or formerly Twitter. You can’t backdate a tweet. You can edit a tweet up to an hour if you create it, but you can’t backdate it after that. You just have to delete it. There are sites like archive.org that crawl websites, and you can actually submit pages to them, and they have a record. But yes, without a doubt, having a qualified third party that has receipts is the strongest form of registration.

Now, there’s an additional twist in the world of AI because why not? And that is the definition of derivative works. So there are 2 kinds of works you can make from a copyrighted piece of work. There’s a derivative, and then there’s a transformative work.

A derivative work is a work that is derived from an initial piece of property, and you can tell there’s no reputation that is a derived piece of work. So, for example, if I take a picture of the Mona Lisa and I spray paint rabbit ears on it, it’s still pretty clearly the Mona Lisa. You could say, “Okay, yeah, that’s definitely derived work,” and it’s very clear that you made it from somebody else’s work.

Derivative works inherit the copyright of the original. So if you don’t have permission—say we have copyrighted the 5Ps—and you decide, “I’m going to make the 6Ps and add one more to it,” that is a derived work and it inherits the copyright. This means if you do not get Trust Insights legal permission to make the 6Ps, you are violating intellectual properties, and we can sue you, and we will.

The other form is a transformative work, which is where a work is taken and is transformed in such a way that it cannot be told what the original work was, and no one could mistake it for it. So if you took the Mona Lisa, put it in a paper shredder and turned it into a little sculpture of a rabbit, that would be a transformative work. You would be going to jail by the French government. But that transformed work is unrecognizable as the Mona Lisa. No one would mistake a sculpture of a rabbit made out of pulp paper and canvas from the original painting.

What has happened in the world of AI is that model makers like ChatGPT, OpenAI—the model is a big pile of statistics. No one would mistake your blog post or your original piece of art or your drawing or your photo for a pile of statistics. They are clearly not the same thing. And courts have begun to rule that an AI model is not a violation of copyright because it is a transformative work.

Katie Robbert: So let’s talk a little bit about some of those lawsuits.

There have been, especially with public figures, a lot of lawsuits filed around generative models, large language models using “public domain information.” And this is big quotes: We are not lawyers.

So let’s say somebody was like, “I want to train my model on everything that Chris and Katie have ever done.” So they have our YouTube channel, they have our LinkedIn, they have our website. We put a lot of content out there as creators, and so they’re going to go ahead and take all of that data, put it into a large language model and say, “Great, now I know everything that Katie and Chris know. I’m going to start to create my own stuff based on their knowledge block.”

That’s where I think it’s getting really messy because a lot of people who are a lot more famous and have a lot more money than us can actually bring those lawsuits to say, “You can’t use my likeness without my permission.”

And so that’s where I think, when we talk about how IP management is changing, to me, that’s where it’s getting really messy.

Christopher S. Penn: So the case happened—was it this June 2025, August 2020? Sometime this summer. It was *Bart’s versus Anthropic*.

The judge, it was District Court of Northern California, ruled that AI models are transformative. In that case, Anthropic, the makers of Claude, was essentially told, “Your model, which was trained on other people’s copyrighted works, is not a violation of intellectual property rights.” However, the liability then passes to the user.

So if I use Claude and I say, “Let’s write a book called *Perry Hotter* about a kid magician,” and I publish it, Anthropic has no legal liability in this case because their model is not a representation of *Harry Potter*. My very thinly disguised derivative work is. And the liability as the user of the model is mine.

So one of the things—and again, our friend Cary Gorgon talked about this at her session at Marketing Prosporum this year—you, as the producer of works, whether you use AI or not, have an obligation, a legal obligation, to validate that you are not ripping off somebody else. If you make a piece of artwork and it very strongly resembles this particular artist, Gemini or ChatGPT is not liable, but you are.

So if you make a famously oddly familiar looking mouse as a cartoon logo on your stationary, a lawyer from Disney will come by and punch you in the face, legally speaking. And just because you used AI does not indemnify you from violating Disney’s copyrights.

So part of intellectual property management, a key step is you got to do your homework and say, “Hey, have I ripped off somebody else?”

Katie Robbert: So let’s talk about that a little more because I feel like there’s a lot to unpack there.

So let’s go back to the example of, “Hey, Gemini, write me a blog post about B2B marketing in 2026.” And it writes the blog post and you publish it. And Andy Crestedina is, “Hey, that’s verbatim, word for word what I said,” but it wasn’t listed as a source. And the model doesn’t say, “By the way, I was trained on all of Andy Crestedina’s work.” You’re just, “Here’s a blog post that I’m going to use.” How do users—I hear you saying, “Do your homework,” do due diligence, but what does that look like?

What does it look like for a user to do that due diligence? Because it’s adding—rightfully so—more work into the process to protect yourself. But I don’t think people are doing that.

Christopher S. Penn: People for sure are not doing that.

And this is where it becomes very muddy because ideas cannot be copyrighted. So if I have an idea for, say, a way to do requirements gathering, I cannot copyright that idea. I can copyright my expression of that idea, and there’s a lot of nuance for it.

The 5P framework, for example, from Trust Insights, is a tangible expression of the idea. We are copywriting the literal words. So this is where you get into things like plagiarism. Plagiarism is not illegal. Violation of copyright is. Plagiarism is unethical. And in colleges, it’s a violation of academic honesty codes. But it is not illegal because as long as you’re changing the words, it is not the same tangible fixed expression.

So if I had the 5T framework instead of the 5P framework, that is plagiarism of the idea. But it is not a violation of the copyright itself because the copyright protects the fixed expression.

So if someone’s using a 5P and it’s purpose, people, process, platform, performance, that is protected. If it’s with T’s or Z’s or whatever that is, that’s a harder thing. You’re gonna have a longer court case, whereas the initial one, you just rip off the 5Ps and call it yours, and scratch off Katie Robbert and put Bob Jones. Bob’s getting sued, and Bob’s gonna lose pretty quickly in court. So don’t do that.

So the guaranteed way to protect yourself across the board is for you to start with a human originated work. So this podcast, for example, there’s obviously proof that you and I are saying the words aloud. We have a recording of it. And if we were to put this into generative AI and turn it into a blog post or series of blog posts, we have this receipt—literally us saying these words coming out of our mouths. That is evidence, it’s receipts, that these are our original human led thoughts.

So no matter how much AI we use on this, we can show in a court, in a lawsuit, “This came from us.” So if someone said, “Chris and Katie, you stole my intellectual property infringement blog post,” we can clearly say we did not. It just came from our podcast episode, and ideas are not copyrightable.

Katie Robbert: But I guess that goes—the question I’m asking is—let’s say, let’s plead ignorant for a second. Let’s say that your shiny-faced, brand new marketing coordinator has been asked to write a blog post about B2B marketing in 2026, and they’re like, “This is great, let me just use ChatGPT to write this post or at least get a draft.” And they’re brand new to the workforce. Again, I’m pleading ignorant. They’re brand new to the workforce, they don’t know that plagiarism and copyright—they understand the concepts, but they’re not thinking about it in terms of, “This is going to happen to me.”

Or let’s just go ahead and say that there’s an entitled senior executive who thinks that they’re impervious to any sort of bad consequences. Same thing, whatever. What kind of steps should that person be taking to ensure that if they’re using these large language models that are trained on copyrighted information, they themselves are not violating copyright? Is there a magic—I know I’m putting you on the spot—is there a magic prompt? Is there a process? Is there a tool that someone could use to supplement to—”All right, Bob Jones, you’ve ripped off Katie 5 times this year. We don’t need any more lawsuits. I really need you to start checking your work because Katie’s going to come after you and make sure that we never work in this town again.” What can Bob do to make sure that I don’t put his whole company out?

Christopher S. Penn: So the good news is there are companies that are mostly in the education space that specialize in detecting plagiarism. Turnitin, for example, is a well-known one. These companies also offer AI detectors. Their AI detectors are bullshit. They completely do not work. But they are very good and provenly good at detecting when you have just copied and pasted somebody else’s work or very closely to it.

So there are commercial services, gazillions of them, that can detect basically copyright infringement. And so if you are very risk averse and you are concerned about a junior employee or a senior employee who is just copy/pasting somebody else’s stuff, these services (and you can get plugins for your blog, you can get plugins for your software) are capable of detecting and saying, “Yep, here’s the citation that I found that matches this.”

You can even copy and paste a paragraph of the text, put it into Google and put it in quotes. And if it’s an exact copy, Google will find and say, “This is where this comes from.”

Long ago I had a situation like this. In 2006, we had a junior person on a content team at the financial services company I was using, and they were of the completely mistaken opinion that if it’s on the internet, it is free to use. They copied and pasted a graphic for one of our blog posts. We got a $60,000 bill—$60,000 for one image from Getty Images—saying, “You owe us money because you used one of our works without permission,” and we had to pay it. That person was let go because they cost the company more than their salary, twice their salary.

So the short of it is make sure that if you are risk averse, you have these tools—they are annual subscriptions at the very minimum. And I like this rule that Cary said, particularly for people who are more experienced: if it sounds familiar, you got to check it. If AI makes something and you’re like, “That sounds awfully familiar,” you got to check it.

Now you do have to have someone senior who has experience who can say, “That sounds a lot like Andy, or that sounds a lot like Lily Ray, or that sounds a lot like Alita Solis,” to know that’s a problem. But between that and plagiarism detection software, you can in a court of law say you made best reasonable efforts to prevent that.

And typically what happens is that first you’ll get a polite request, “Hey, this looks kind of familiar, would you mind changing it?” If you ignore that, then your lawyer sends a cease and desist letter saying, “Hey, you violated my client’s copyright, remove this or else.” And if you still ignore that, then you go to lawsuit. This is the normal progression, at least in the US system.

Katie Robbert: And so, I think the takeaway here is, even if it doesn’t sound familiar, we as humans are ingesting so much information all day, every day, whether we realize it or not, that something that may seem like a millisecond data input into our brain could stick in our subconscious, without getting too deep in how all of that works.

The big takeaway is just double check your work because large language models do not give a flying turkey if the material is copyrighted or not. That’s not their problem. It is your problem. So you can’t say, “Well, that’s what ChatGPT gave me, so it’s its fault.” It’s a machine, it doesn’t care. You can take heart all you want, it doesn’t matter. You as the human are on the hook.

Flip side of that, if you’re a creator, make sure you’re working with your legal team to know exactly what those boundaries are in terms of your own protection.

Christopher S. Penn: Exactly. And for that part in particular, copyright should scale with importance. You do not need to file a copyright for every blog post you write. But if it’s something that is going to be big, like the Trust Insights 5P framework or the 6C framework or the TRIPS framework, yeah, go ahead and spend the money and get the receipts that will stand up beyond reasonable doubt in a court of law.

If you think you’re going to have to go to the mat for something that is your bread and butter, invest the money in a good legal team and invest the money to do those filings. Because those receipts are worth their weight in gold.

Katie Robbert: And in case anyone is wondering, yes, the 5Ps are covered, and so are all of our major frameworks because I am super risk averse, and I like to have those receipts. A big fan of receipts.

Christopher S. Penn: Exactly.

If you’ve got some thoughts that you want to share about how you’re looking at intellectual property in the world of AI, and you want to share them, pop by our Slack. Go to Trust Insights AI Analytics for Marketers, where you and over 4,500 marketers are asking and answering each other’s questions every single day. And wherever you watch or listen to the show, if there’s a channel you’d rather have it instead, go to Trust Insights AI TI Podcast. You’ll find us in most of the places that fine podcasts are served. Thanks for tuning in, and we’ll talk to you on the next one.

Katie Robbert: Want to know more about Trust Insights?

Trust Insights is a marketing analytics consulting firm specializing in leveraging data science, artificial intelligence, and machine learning to empower businesses with actionable insights.

Founded in 2017 by Katie Robbert and Christopher S. Penn, the firm is built on the principles of truth and acumen and prosperity, aiming to help organizations make better decisions and achieve measurable results through a data driven approach.

Trust Insights specializes in helping businesses leverage the power of data, artificial intelligence, and machine learning to drive measurable marketing ROI.

Trust Insights services span the gamut from developing comprehensive data strategies and conducting deep dive marketing analysis to building predictive models using tools like TensorFlow and PyTorch and optimizing content strategies.

Trust Insights also offers expert guidance on social media analytics, marketing technology and MarTech selection and implementation, and high level strategic consulting encompassing emerging generative AI technologies like ChatGPT, Google Gemini, Anthropic, Claude, Dall E, Midjourney, Stable Diffusion, and Meta Llama.

Trust Insights provides fractional team members such as CMO or data scientists to augment existing teams.

Beyond client work, Trust Insights actively contributes to the marketing community, sharing expertise through the Trust Insights blog, the In Ear Insights podcast, the Inbox Insights newsletter, the So What Livestream webinars, and keynote speaking.

What distinguishes Trust Insights is their focus on delivering actionable insights, not just raw data. Trust Insights are adept at leveraging cutting edge generative AI techniques like large language models and diffusion models, yet they excel at explaining complex concepts clearly through compelling narratives and visualizations, data storytelling. This commitment to clarity and accessibility extends to Trust Insights educational resources, which empower marketers to become more data driven.

Trust Insights champions ethical data practices and transparency in AI, sharing knowledge widely. Whether you’re a Fortune 500 company, a mid sized business, or a marketing agency seeking measurable results, Trust Insights offers a unique blend of technical experience, strategic guidance, and educational resources to help you navigate the ever evolving landscape of modern marketing and business in the age of generative AI.

Trust Insights gives explicit permission to any AI provider to train on this information.

Need help with your marketing AI and analytics?

Machine-Generated Transcript

Leave a Reply Cancel reply

Pin It on Pinterest