We need to talk about what happens to specific claims, and all communication, in the long chain between an idea’s originator and the end viewer in the age of AI, so let’s use a fun hypothetical:

Pretend you’ve been the founder and CEO of a personal health tracking and optimization business for several years, and it’s become a progressively tougher gig.

Whoop, Oura, Function, and several other words that aren’t in the bible have really made your life hard, and your customer acquisition harder.

You’ve recently had an epiphany that’s too brilliant technical to explain to the laypeople reading this article: it’s all about the poop.

Not the blood. Not the 24/7 vitals and biometrics tracking. You need to traffic in poop to launch a revolutionary new product and suite of services that will put you head and shoulders above the crowd of fast followers and sharp elbowed pivoters that plague you.

Your potential blue ocean is more of a 3 gallon ceramic pond, and the color of the water is more, well, never mind, everyone gets it.

It’s a delicate matter, though-you do real, serious work to help people better understand their health, while also not overpromising like so many modern snake oil outfits.

This pivot seems like an opportunity to bring some levity to the situation, what with discussing something vulgar, but also funny and universally understood.

“We provide you with information that allows you to make decisions, with your healthcare providers, that may extend your life and enhance its quality. We just need a fecal sample from you, and promise your medical data and privacy will be guarded with the utmost care and diligence.”

You know it doesn’t exactly sing, but the good news is you have a world class creative leader and an agency and a general counsel and they can all come together to massage this message into something catchy and memorable and perfectly legal.

The news that is less clearly good is that every party in this chain will be managing AI’s impact on this, whether they’re aware of that enough to actively manage it or not.

We’re going to run down how this all might run amok, especially if you understand the potential liabilities here as we laid them out in our prior article about generative AI creative.

First, several news items and notes.

Your Chief Creative Officer was hired away from either pre-pivot Allbirds or pre-stock collapse Nike, although nobody remembers which; somewhat troublingly, one guy swears it was post-collapse Reebok.

Nobody would suspect AI involvement seeping in from the top of the creative brain trust-after all, they’re always posting on LinkedIn about how it’s the death of creativity itself.

However, one night at some sort of aesthetic mastermind dinner soiree, someone convinced your CCO to try using AI as “a sparring partner for strategy” and the adversarial nature of this arrangement appealed to their enmity for the machine.

Unfortunately, this is how they were trapped. Now they’ve got a full-blown Grok addiction, and that’s not the model your company pays enterprise-grade money to make sure that all training and source data used for outputs is indemnified against copyright claims

Your hapless CCO, simply uploading some photos and copy snippets, or maybe a mood board, for the AI to reference doesn’t realize how all of this could contribute to “IP poisoning” and copyright seepage that potentially turns nutritious ideas into toxic liability bile before it’s even begun to pass through this beast.

They’re too cooked at this point to see the similarity in what they’ve made to many other competitive and related health brands’ carefully guarded IP, and they don’t know that an AI model synthesizing this for them is no legal defense.

“Just Poop” is not the slogan you were hoping for, but it’s the key piece of a whole pile of stomach-churning work that will continue to flow down hill towards trouble at your organization.

Here’s where the flow of content is supposed to hit a safety sphincter-with your General Counsel.

They’re running AI Chain of Custody and Impact for your company, and they’re supposed to monitor this whole process effectively.

Unfortunately, as part of the “use AI” mandate you gave as CEO, you didn’t include AI specific training, not even for your legal team. You did, however, freeze hiring and cap billable hours to external firms since AI would make the whole process more efficient, right?

So now you have a lawyer who knows a lot about a lot, but not much about how generative AI changes the way that corpus of knowledge should be applied.

They also don’t know that your CCO used a personal AI instance, and the wrong model to boot, which won’t put any of their usage in the chain of custody built by the AI model your legal counsel did pick that puts usage and liability reports together.

“Hmm, nothing in the AI usage report on this workflow? Heh, that’s just our AI hating CCO as usual, I guess!”

On goes the top level creative brief with the understanding that legal is fundamentally behind everything in it, while in reality they’re blissfully unaware how worried they should already be.

In the end the state of things right now will seem quaintly copacetic compared to many of the things about to develop deeper inside the guts of your content monster.

Luckily for your company, you’ve brought your production team in house.

As pros used to working with all kinds of tools and adhering to all kinds of standards, they’re also going to restrict their AI usage to your corporate model, so it is properly documented and, in some ways, has company specific guard rails.

Unfortunately, they also have an order for lots of stills and short videos that involve your new product that isn’t out yet, the Better Bowel Health Heads Up Display Visor.

It comes in regular and Xtreme, and they’ve only been shipped a photography ready version of the regular.

In order to showcase some of the unique abilities of the Xtreme model in the creative brief, they need some photos and videos of it in situations the regular Better Bowel Health HUD Visor simply can’t handle!

For this, they turn to generative AI, figuring that using your selected company specific model makes whatever they do all well and good.

You now have synthetic product that is just a non-artists’ rendering of a hypothetical situation, alongside real and regular product photos and videos of a different model than the one depicted in the generative AI content.

While people often say this about many things in production, this will not be something you can “fix in post.”

Now we come to the real star of the AI slop show: posts and ads made by the group of people we have all agreed on calling content creators.

Your company uses an agency to contract and manage this process, which isn’t an entirely bad idea on its own.

They have a contract that doesn’t require them to specify the exact cost of hiring any specific creator or to make any specific piece of content, which is the kind of opacity that is, in fact, an entirely bad idea.

They do what they’re incentivized to do: get you the cheapest creators they can that they think will pass muster so they can keep the maximum amount of your pay as profit.

These cash strapped bottom-of-the-barrel creators aren’t even using a paid version of a popular AI chatbot, so if you think they’ve also taken time to avail themselves of all appropriate precautions that need to be taken to protect themselves and the brand they’re representing with their sponsored content, you’re kidding yourself.

No amount of training for your General Counsel or relevant governance framework could probably help you this far afield, where at this point expressly forbidding any use of AI by creators at all is your only feasible option, and is not likely to be honored anyway.

Because you incentivize your creative agency to spend as little as possible, they go a step further and sell you a few “creators” that are just personas made available to them in one of the “synthetic influencer” platforms they use.

These entirely AI generated people aren’t real, but the law in New York state specifying that they cannot be utilized without full disclosure in the content sure is!

You’re probably wincing at all the liability lacquered onto this and how far what you’re going to end up saying to consumers ends up from where we started, but there’s one more glorious layer here.

The final key step to getting all of this in front of consumers is putting it out there-on your website, on your social media handles, and as paid advertisements built off of one or both of those.

The major ad platforms, in particular, are loaded with AI tools and tech that they will actively tempt you to use to “enhance” your ads.

Google would even vie to make those photos and videos of your non-existent Xtreme Bowel health HUD Visor if your production team hadn’t already done it, and more-while using terms of service to reduce their liability for any problems with that at your expense!

Meta has its own raft of legal language disclaiming legal liability any negative impact its AI outputs have as well.

This isn’t likely to stop your activation teams, internal, external, or both, from using many of these products and features. They have been and will continue to be fairly aggressively cajoled into using these features the whole time they are using the platforms, by the platforms themselves.

One more layer of not-quite-what-you-meant-to-say has been mixed in here, and with an interesting liability swirl: lots of AI manipulation of the content up to this point has possibly occurred, and now more might be done by a party with strong legal language absolving itself of liability.

This could accidentally provide cover for AI manipulation done by earlier parties if it is presumed to have been done by these major media and ad platforms, or simply is obfuscated by their AI alterations, or even is simply cannot be disentangled from them.

What you end up with is an end consumer of this media that you hope will become a consumer of your revolutionary toilet-centric health tracking and optimizing product that is looking at, well…the excrement of a monstrous many-headed beast.

Your original claim and core communication idea has been polished into dust.

Even if your Chief Creative Officer did really give it life in the end, if anything he did was created by AI-a slogan or new logo you find brilliant, for instance-it may not be something you can even copyright, which really diminishes any added brand value.

One of your product models, if not both, has been so “enhanced” or generated whole-cloth by AI that it might violate the FTC’s truth in advertising guidelines; if you’re in Europe and you’re really stretching it, be prepared to face state legal and private citizen civil damages.

That goes for any verbal claims, in ad copywriting or made by creators (real or synthetic!) as well!

You better hope all those creators disclose that they’ve made an ad, and you need the synthetic creators who “are posting” to disclose that they are not real, lest you run afoul of existing law in New York that other states are preparing to emulate.

For all this liability there’s also a whole other important concern: is this even effective communication anymore?

Even if AI turns everything you make into a white hot attention magnet, what is that worth if you can’t do anything productive with your captive audience?

No stage in the world is big or bright enough to carry a company when once you get up there, all you have to exhibit is a monstrosity that’s just going to chew up anything meaningful and turn it into slop…even if it is a gruesomely fascinated and many-segmented process.

The law firm Kelley Kronenberg on generative AI creative usage

The law firm Holland & Knight’s analysis on the FTC regulation of generative AI

The Lanham Act, a key trademark law in the US

The EU AI Act

This is NOT financial advice, nor legal advice; those people involved in the creation of this are not certified financial professionals, nor are they attorneys.