Coca-Cola’s AI Christmas Ad: What the Backlash Tells Us

The Coca-Cola AI ad controversy is less a story about artificial intelligence and more a story about expectations. When Coca-Cola released its AI-generated remake of its iconic 1995 “Holidays Are Coming” Christmas ad, the public reaction was swift and largely negative. Critics called it soulless, uncanny, and a cynical cost-cutting exercise dressed up as innovation.

What makes it worth examining is not whether the ad was good or bad. It is what the response reveals about where brands, agencies, and audiences actually stand on AI-generated creative in 2025.

I have spent more than 20 years in agency leadership, and I have watched a lot of creative decisions get made for the wrong reasons. Budget pressure dressed up as innovation. Efficiency framed as ambition. The Coca-Cola situation has echoes of both, and it is worth being honest about that rather than simply picking a side in the AI debate.

What Coca-Cola Actually Did

To understand the controversy, you need to understand what was being remade. The original “Holidays Are Coming” ad from 1995 is one of the most recognisable pieces of Christmas advertising ever made. The image of illuminated Coca-Cola trucks rolling through a snow-covered landscape became genuinely iconic. It has been rebroadcast year after year for three decades. People have emotional associations with it that go well beyond the brand itself.

Coca-Cola commissioned an AI-generated version of that ad using generative video tools. The result preserved the broad visual concept but introduced the kind of artefacts and inconsistencies that current generative video technology still struggles to eliminate. Hands that do not look quite right. Motion that feels slightly off. A general sense that something is missing, even when it is hard to name exactly what.

The brand positioned it as a celebration of innovation. A significant portion of the audience experienced it as a diminishment of something they cared about.

If you want a broader grounding in how AI is being applied across marketing right now, the AI Marketing hub at The Marketing Juice covers the landscape without the hype. This controversy sits inside a much larger shift, and understanding the context matters.

Why the Backlash Was So Strong

There are three reasons the reaction was sharper than it might have been for a different brand or a different ad.

First, nostalgia is a fragile material. The original ad works because it carries thirty years of accumulated emotional association. People do not just see trucks and snow. They see their own memories of Christmas mornings and childhood excitement and the specific feeling of a particular era. When you attempt to recreate that using a process that, by definition, has no access to any of those associations, you are working against yourself from the start.

Second, the uncanny valley problem is considerably more acute in video than in still images or text. I have seen AI-generated copy and AI-assisted imagery integrated into campaigns with very little audience friction. But generative video is still at a stage where human perception picks up on something being wrong even when the viewer cannot articulate what it is. The motion is slightly too smooth or slightly too jerky. Faces do not quite hold. The physics of light behaves in ways that are almost right but not quite. Audiences feel this before they think it.

Third, and perhaps most importantly, people suspected the motivation. When a brand with Coca-Cola’s resources releases an AI-generated version of one of its most beloved ads, the immediate assumption is that this was a cost decision. Whether that is true or not almost does not matter. The perception that a corporation chose to cut corners on something culturally significant is a reputationally expensive position to be in.

The Brief Was Wrong Before the Technology Was Chosen

I spent several years judging the Effie Awards, which are specifically focused on marketing effectiveness. One thing that experience reinforced is how often creative failures trace back to a brief problem rather than an execution problem. The work goes wrong long before anyone opens a production tool.

The Coca-Cola AI ad looks, to me, like a brief problem. The decision to recreate an existing iconic ad using generative AI is a brief that was almost certainly going to produce something that disappointed. Not because AI cannot generate compelling video. It can, and it will get better at it. But because the specific task of faithfully recreating a beloved nostalgic asset is precisely the kind of job where the gap between “good enough” and “as good as the original” is most visible and most costly.

Early in my career, I learned a version of this lesson the hard way. I was working on a campaign where the instinct was to take a previous year’s successful approach and replicate it with a smaller budget. The logic seemed sound. Do what worked before, just more efficiently. What we underestimated was how much of the original’s success came from the craft and the production quality, neither of which could be replicated at a reduced spend. The campaign landed flat. The brief had set us up to fail.

The question Coca-Cola’s team should have been asking is not “can we use AI to remake this ad?” but “what is the right job for AI in our Christmas marketing this year?” Those are different questions with different answers.

What AI Is and Is Not Good at in Creative Production

The controversy has generated a lot of polarised commentary. Some people have used it as evidence that AI has no place in creative marketing. Others have dismissed the critics as technophobes who will come around eventually. Neither position is particularly useful.

AI tools are genuinely strong in a specific set of creative applications. Rapid concept generation and iteration. Producing variations at scale. Supporting teams that are under-resourced. Handling production tasks that are time-consuming but not craft-dependent. Generating first drafts that human editors then shape into something better. These are real, commercially valuable capabilities, and the Semrush overview of AI in marketing gives a reasonable sense of how these applications are being mapped across different functions.

Where AI is currently weak, and where the Coca-Cola ad ran into trouble, is in tasks that require emotional fidelity to something that already exists in people’s memories. Generative AI works from patterns in training data. It does not know what the original “Holidays Are Coming” ad meant to the people who watched it in 1995. It cannot replicate that meaning because it has no access to it. It can only approximate the visual surface.

This is not a criticism of the technology. It is an accurate description of what the technology is. The problem is when a brief asks it to do something it cannot do, and then the result is judged against a standard it was never equipped to meet.

At iProspect, when I was leading the agency through a significant growth period, we had a standing principle about tool selection: the tool should follow the brief, not the other way around. It sounds obvious. In practice, when a new technology is generating excitement internally, the temptation is to find briefs that fit the tool rather than tools that fit the brief. That inversion is where a lot of expensive mistakes get made.

The Cost-Cutting Perception Problem

There is a broader issue here that goes beyond this specific ad. When large, profitable brands use AI-generated content in high-profile campaigns, the public interpretation is almost always that it is about saving money. Sometimes that interpretation is correct. Sometimes it is not. But the perception exists, and brands need to be clear-eyed about it.

Coca-Cola is one of the most valuable brands on earth. It is not a company that needs to cut the production budget on its Christmas ad. The fact that it chose to use generative AI for a high-profile creative asset was always going to invite scrutiny about motivation, regardless of the quality of the output.

Compare this to a challenger brand using AI to produce creative at a scale it otherwise could not afford. The same technology, a completely different narrative. One reads as innovation born of necessity. The other reads as cost-cutting dressed up as progress.

Brand context shapes how every creative decision is interpreted. This is not new. What is new is that AI-generated content carries a specific set of associations right now, and brands need to factor those associations into their decisions about when and how to use it publicly.

What the Creative Industry Got Right and Wrong in Its Response

The creative industry’s response to the Coca-Cola ad was, in some quarters, more theatrical than useful. There was a significant amount of commentary that used the ad as a proxy for a much larger argument about AI replacing human creativity, which is a legitimate debate but not quite the debate the ad itself raises.

The more grounded criticism, and the one worth taking seriously, is about craft and intentionality. Good advertising, at its best, is made by people who understand what they are trying to make the audience feel and why. That understanding shapes thousands of small decisions across a production. The choice of a specific shade of light. The precise timing of a musical cue. The way a character’s expression changes in a single frame. These things are not accidental in great work. They are the product of human judgment applied repeatedly across a long production process.

Generative AI does not replicate that process. It produces outputs based on pattern recognition. For many applications, that is entirely sufficient. For an ad that is supposed to carry thirty years of emotional weight, it is not.

What the creative industry got wrong is the leap from “this specific application was misjudged” to “AI has no place in creative production.” Those are very different claims. The tools being built right now are changing what is possible in production, in ideation, in personalisation at scale. Dismissing all of that because one high-profile execution landed badly is not critical thinking. It is a defensive reaction.

There is a lot more nuance to work through across the AI and marketing landscape. The AI Marketing section at The Marketing Juice covers the tools, the workflows, and the honest limitations in more depth, if you want to go further than the headlines.

What This Means for Marketing Teams Right Now

If you are leading a marketing team and trying to figure out how to position your use of AI in creative work, the Coca-Cola controversy offers a few practical lessons.

Be honest about motivation. If you are using AI because it reduces cost, that is a legitimate business reason. But do not dress it up as innovation if it is not. Audiences and industry observers are getting better at reading the difference, and being caught in that gap is worse than being straightforward about the trade-offs you are making.

Match the tool to the brief, not the brief to the tool. AI-generated video is improving rapidly. It is already genuinely useful for certain types of production. But it is not yet at the stage where it can faithfully recreate something that carries deep emotional associations without visible artefacts. Know what the technology can and cannot do before you commit to a brief that depends on it.

Consider the audience’s frame of reference. The people watching the Coca-Cola ad were not evaluating it as a technical achievement. They were comparing it to something they already loved. That is an almost impossible standard to meet with any technology, AI or otherwise. If your creative brief is essentially “do what we did before but differently,” you need to think very carefully about whether that is a brief worth pursuing.

Think about where AI actually adds value in your production process rather than where it is most visible. Some of the most effective uses of AI in creative production are invisible to the audience. Faster iteration. More options tested in pre-production. Better personalisation in distribution. These are commercially valuable applications that do not carry the same reputational risk as a high-profile AI-generated hero film.

The Moz Whiteboard Friday on generative AI for content is worth watching if you are thinking through how AI fits into your content production workflow. It is grounded and practical rather than evangelical, which is the register most marketing teams actually need right now.

The Longer View

In five years, the Coca-Cola AI ad will probably be a footnote in a much larger story about how generative AI was integrated into creative production. The technology is improving at a pace that makes today’s limitations look temporary, because most of them are.

What will not change is the underlying principle that the brief matters more than the tool. I have seen this across every significant technological shift in the industry. When paid search arrived, the teams that did well were not the ones who were fastest to spend money on clicks. They were the ones who understood what they were actually trying to achieve and built campaigns around that. When social media arrived, the brands that built durable audiences were not the ones who posted most frequently. They were the ones who had something worth saying.

I remember running a paid search campaign at lastminute.com for a music festival. The campaign was not technically sophisticated. But the brief was sharp. We knew exactly who we were talking to, what they cared about, and what we needed them to do. The revenue followed. The technology was almost incidental to the clarity of the thinking behind it.

AI will follow the same pattern. The teams that get the most from it will be the ones who are clearest about what they are trying to achieve before they open any tool. The teams that struggle will be the ones who let the technology drive the brief rather than the other way around.

The Coca-Cola controversy is a useful data point. It shows where the current technology falls short and what the reputational risks look like when a high-profile execution does not land. But it is not a verdict on AI in creative marketing. It is one expensive lesson in brief-setting, and the industry will learn from it.

The Ahrefs AI tools webinar series is worth bookmarking if you want to track how practitioners are actually integrating these tools into their workflows, as opposed to how vendors are selling them. The gap between those two things is still considerable.

About the Author

Keith Lacy is a marketing strategist and former agency CEO with 20+ years of experience across agency leadership, performance marketing, and commercial strategy. He writes The Marketing Juice to cut through the noise and share what works.

Frequently Asked Questions