Let me tell you what Coca-Cola did.

In November 2024, the company that once hired the best human beings in advertising to craft one of the most beloved Christmas images in commercial history -- the red trucks rolling through the snow, the warmth, the feeling, the thing that made people actually stop and watch a commercial -- decided instead to feed that entire legacy into a computer and see what came out.

Let me be more precise about what actually happened here. Five AI specialists worked for approximately one month. They generated roughly 70,000 video clips. One hundred Coca-Cola staff were involved across the broader production. After all of that, after all those people and all those compute hours and all that corporate enthusiasm for the future of AI-generated content, what came out the other end was described by actual viewers as "soulless," "creepy," "lifeless," and "digital slop."

The AI-generated animals looked, per one critic, "part shiny, part plastic." The trucks changed shape between frames. The characters had the dead-eyed stare of something that had never actually seen a human face and was doing its absolute best to approximate one from a training dataset. It could not pull it off.

Coca-Cola's response to this very public, very documented, very-covered-by-every-trade-publication catastrophe?

Do it again in 2026. Same AI tools. More animals. Slightly improved truck wheel animations. Still fundamentally wrong in the way that only something built by a machine pretending to understand human warmth can be fundamentally wrong.

Alex Hirsch, creator of Gravity Falls and a person who has clearly had enough of watching this industry eat itself, posted what the entire creative community was thinking: "FUN FACT: @CocaCola is 'red' because it's made from the blood of out-of-work artists." The company's official response was essentially: we tested it privately with real consumers and they liked it fine, actually. The genie is out of the bottle. Progress marches forward. Please enjoy your digitally rendered holiday warmth and stop asking uncomfortable questions about the six-fingered Santa.

This is advertising in 2026.

It is capitalism dressed up as creativity. And they are not even bothering to dress it up very convincingly anymore.

Coca-Cola is nowhere near alone in the embarrassment and the industry knows it. McDonald's Netherlands dropped an AI-generated holiday ad so cheerless, so visually incoherent, and so thoroughly missing the emotional register of what holiday advertising is supposed to do, that the comments section declared it had "ruined Christmas spirit" in a way that was not, in context, hyperbole. The company quietly pulled it with no announcement, no apology, and no public explanation whatsoever. One day the video link worked. Then it did not.

Paramount Pictures promoted its film Novocaine with AI scripting and narration so robotic in delivery and so devoid of personality that observers noted it was functionally indistinguishable from the spam content farms generating low-quality YouTube videos at 3 AM. Skechers ran AI-generated imagery in actual physical print magazines, including in Vogue, of all publications on earth, featuring human hands with the wrong number of fingers and anatomical arrangements no licensed physician would recognize as occurring in nature. "The determination not to use a human illustrator is incredible," one person wrote in response. "The fact that they sent it to print with such noticeable AI errors is CRAZY," wrote another. Both of these people are correct.

And then it keeps going. H&M announced it was building digital twins of its human models, meaning copies of real people's faces and bodies generated by AI, specifically to avoid the considerable inconvenience of paying those real people to show up and be photographed. Valentino produced an AI luxury campaign that the fashion press, an industry not renowned for its willingness to criticize luxury brands, described as "cheap" and "tacky." Friend AI put up subway posters that were vandalized by commuters within hours of installation, the public providing its own editorial response in permanent marker. Meta's Advantage+ platform autonomously replaced a clothing brand's top-performing ad creative with an AI-generated image of an elderly woman wearing the brand's products, without asking anyone's permission, generating immediate backlash and the kind of brand safety incident that ends careers. Call of Duty: Black Ops 6 shipped with loading screen artwork featuring a zombified Santa Claus with six fingers on one hand. Six fingers. In a shipped, released, commercially available product that presumably had people looking at it before it went out.

None of this is subtle. None of it is ambiguous. None of it is a mystery. And yet the industry keeps going.

Merriam-Webster named "slop" its official word of the year for 2025. This is the linguistic establishment formally acknowledging that we have, as a society, collectively named and accepted into the vocabulary a brand new category of digital garbage. AI-generated content, mass-produced at industrial scale, quality-inspected by no one, optimized entirely for cost reduction and volume output rather than the small, annoying, inconvenient thing where it actually works on human beings.

The scale of the problem is genuinely staggering once you start looking at the numbers. On Quora and Medium, AI-generated material jumped from under 2% of all content in 2022 to nearly 40% of all content by 2024. Roughly 21% of all YouTube content now qualifies as AI slop by any reasonable definition. Integral Ad Science identified the worst of the AI slop sites as one of the most critical threats to digital advertising effectiveness, and found that over 90% of known AI-generated garbage sites remained unlisted on the major ad-buying platforms' blocklists. What that means in plain language is that brands are paying real money to run their advertisements on digital garbage dumps, and their agencies are describing this inventory as premium placement.

The consumer trust numbers are, predictably, an absolute disaster for everyone involved in this enterprise. A 2025 study from the Nuremberg Institute for Market Decisions found that simply labeling an ad as AI-generated makes people perceive it as less natural, less useful, and less credible before they have watched a single second of it. The label alone is enough to poison the well. Researchers at Washington State University, publishing in the Journal of Hospitality Marketing and Management, found that the presence of AI in product descriptions simultaneously drops both emotional trust and purchase intention in the same consumers at the same time. 36% of American adults say they are less likely to buy from a brand that uses AI in its advertising, a number that has moved four points in the wrong direction in just a few months. Only 10% say they are more likely. The gap between how the marketing industry feels about AI (77% positive, according to research) and how the consumers that marketing industry is trying to reach feel about AI (38% positive) is not a nuance or a rounding error. It is a chasm that the industry is actively making wider every quarter.

"Consumer trust is reduced when ads are clearly labeled as AI-generated, especially in emotionally high categories," said Arun Roongta, Managing Director at Texzone Information Services, speaking to Storyboard18. "AI must be a backstage enabler, with the brand's human narrative remaining center-stage."

The industry heard something different. It heard: center stage, robots, now, as fast as possible, fire the humans first.

Former Ogilvy chairman and CEO Ken Roman, a person who has spent considerably more time in this industry than the executives currently dismantling it, asked the question that nobody at the major holding companies actually wants to answer out loud: does AI have fingerspitzengefühl? The German concept, literally meaning sensitivity in the fingertips, refers to the intuitive feel for when something is right, when the work lands, when the human truth has been captured accurately. Does the machine know when it is right? Roman's answer was "Maybe." He followed it immediately with: "Until then, the great ads will be created by brilliant copywriters and art directors."

That single word "maybe" is carrying enormous freight.

Here is what is actually, structurally happening, and why it is going to get worse before it gets better. Brands are cutting production budgets. AI makes it cheaper and faster to produce content. So they produce more content, at lower quality, and flood every available channel with it. The ads do not work as well because the trust is lower. The trust is lower because the quality is lower. So the brands produce more content to compensate for the lower performance, which drives trust down further because there is now more low-quality content. Which leads to lower performance. Which leads to more content. It is a doom loop dressed up in a press release about transformation and innovation and the exciting future of AI-powered creative.

Bob Hoffman, known professionally as The Ad Contrarian and described by his peers as the advertising industry's most reliable truth-teller and most persistent pain in the industry's side, wrote the joke version of all of this years ago, before any of it was happening: "Now that computers can write and design ads, we can get down to the real business of advertising. You know, meetings and downloads and uploads and briefings and off-sites and Powerpoints and metrics."

He was being sardonic.

He was describing the present with startling accuracy. He was kidding.

He was not kidding.

The Coca-Cola autopsy. The Omnicom math. Jay Pattisall's bloodbath forecast. Andrea Brimmer on building a human brand. Alex Schultz admitting AI "will destroy jobs" without the euphemisms. Bob Hoffman's prophecy. The Spotify confession. The whole ugly, documented, sourced collapse of an industry pretending it is having a transformation instead of a crisis.

Free. All of it.

The sourcing behind the sourcing. The names that did not make the final draft because they spoke off the record. The specific numbers from the Forrester briefing that the public summary softened. The Omnicom integration details that trade press buried because Omnicom buys ads in trade press.

The follow-up. What happened after the Dentsu cuts landed. Where the 700 Ogilvy people went. Which holding company is next and why the math says it cannot be avoided. Which CMOs are privately protecting their human creative budgets while publicly performing AI enthusiasm for their boards.

The stuff nobody will print. The junior pipeline data by agency. The real reason Meta's Advantage+ replaced your best ad with an AI grandmother without asking. The adtech deals that explain why your brand safety problem is worse than any vendor will tell you it is.

The people. Conversations with the creative directors, media buyers, and CMOs who know exactly what is happening and will only say it somewhere that is not LinkedIn.

It is the publication for people who work in this industry and are tired of being lied to about it by the trades that need access, the vendors that need coverage, and the holding companies that need the narrative managed.

Pesach Lattin has been covering advertising, adtech, and media longer than most of the industry's current executives have been in it. With the irreverence the industry needs and almost never gets. Without the deference that makes most coverage useless.

The industry is restructuring itself around AI in ways that will determine which brands, agencies, and careers survive the next five years. The people who understand what is actually happening are already subscribers. The people relying on the trades to tell them what the trades' advertisers want them to know are going to be surprised in ways that will not feel good.

You just read the free version and it was better than most of what you pay for.

Imagine the rest.

Subscribe to ADOTAT+ now.