The Association of National Advertisers chose two words of the year for 2025. For the first time in the award's history, they could not pick just one. The words were "authenticity" and "agentic AI." The tension between them is not a coincidence. It is the central problem of marketing in 2026.

Brands are using AI more than ever. A 2024 NIM study of 600 marketing professionals found universal AI adoption within the sample, with efficiency and output quality cited as the main drivers (Buder et al., NIM, 2024). At the same time, consumers are pulling back. Only 26% of consumers prefer generative AI creator content to traditional creator content, down from 60% in 2023. The gap between how enthusiastically brands have adopted the technology and how consumers are responding to it is widening at a rate nobody in a strategy meeting seems to be taking seriously enough (Digiday, 2026).

The obvious response is to say that brands should use AI more carefully. But that framing misses the actual problem. The issue is not that brands are using AI. It is that many of them are using it as a replacement for the kind of thinking that cannot be automated: the judgment about what a brand actually sounds like, what it believes, how it makes people feel, and why any of that should matter to anyone.

The pasteurization problem

There is a specific failure mode that is becoming visible across industries. A brand adopts AI for copywriting, strategy development, asset creation, and campaign versioning. The output is competent. It is grammatically correct, tonally appropriate, and statistically optimised for engagement. It is also indistinguishable from the output of every other brand using the same tools with the same prompts and the same optimisation logic.

This is what I mean by pasteurization. The brand does not disappear. It survives, technically intact, with a logo and a colour palette and a mission statement. But the texture that made it recognisable, the idiosyncrasies, the point of view, the things that could only have come from specific humans with specific experience, gets smoothed out in the process. "I can tell when somebody's used a ChatGPT script," one creator told Digiday. "Part of the issue is that creators are outsourcing their creativity." The same logic applies to brands.

AI optimises for the statistically probable. The statistically probable is, by definition, the average of everything that came before it. A brand that delegates its creative voice entirely to that process is not building an identity. It is building a mean.

The trust penalty

The consumer response to this is already measurable. Research by the Nuremberg Institute for Market Decisions found that even technically polished AI content faces a "trust penalty": simply knowing that a piece of content was crafted by an algorithm makes people trust it less and engage with it less enthusiastically. Transparency alone "reveals a fundamental problem but doesn't solve it."

WARC's own research sharpens this finding: AI ads perceived as human-made had the highest click-through rates, while ads perceived as AI-generated were penalised. The performance gap is not about quality. It is about origin. Consumers are not rejecting AI because the output is bad. They are rejecting it because they can tell something human is missing, and that absence registers as a form of dishonesty (WARC, 2026).

McDonald's pulled its AI-generated Christmas ad. Coca-Cola's AI holiday campaign generated significant backlash despite the brand's scale and budget. Both brands failed to account for what researchers call the "authenticity premium": the intangible but measurable value that human creativity provides in emotionally meaningful contexts. The financial case for AI production may be strong. The risk calculus, once public response is factored in, is considerably less clear.

What AI cannot do

The deeper problem is not technical. It is cognitive. AI does not have a background. It does not have the accumulated experience of working in a specific market, building relationships with a specific audience, failing publicly, recovering, and learning something that cannot be written down. It does not have the emotional memory that shapes what a brand communicates when the stakes are high.

Brand identity, at its core, is a product of human judgment operating under constraint over time. The constraint is the market. The judgment is what decides which signals matter, which compromises to make, which moments to seize and which to let pass. That judgment is not just strategic. It is embodied in the people who have been living with the brand long enough to know what it sounds like when it is being true to itself versus when it is performing a version of itself for an algorithm.

AI can accelerate the execution of that judgment. It cannot replace the judgment itself. When brands forget this, they do not make a technological error. They make an identity error. As Professor Colleen Kirk of the New York Institute of Technology notes, "Consumers are becoming ever more skeptical of the human origin of advertisements and marketing messages. While AI tools offer marketers an exciting new frontier, these professionals should bear in mind a time-tested principle: authenticity is always best."

The lazy use problem

The word "lazy" is doing real work here. Using AI is not lazy. Using AI without asking what it is costing you in terms of distinctiveness, voice, and emotional texture is lazy. The difference is the presence or absence of a human with enough investment in the brand to notice when something feels wrong.

As one industry observer put it: "2026 is going to be all about getting really specific about how you use AI in your workflow." The brands that will navigate this well are the ones that treat AI as infrastructure rather than identity. It handles volume, versioning, and efficiency. The questions of what the brand believes, how it talks to people, and why anyone should care: those remain stubbornly human problems.

The brands losing ground right now are not the ones using AI. They are the ones that stopped asking those questions, and handed the answers to a tool that was never designed to provide them.

‍

Photo by cottonbro studio: https://www.pexels.com/photo/a-woman-with-number-code-on-her-face-while-looking-afar-5473956/

‍

References

• WARC. (2026). Marketers Love AI. Consumers Are Starting to Push Back. WARC Strategy. Link
• Nuremberg Institute for Market Decisions / NIM. (2024). Transparency Without Trust: Consumer Attitudes Toward AI-Generated Marketing Content. NIM. Link
• Digiday. (2026). After an Oversaturation of AI-Generated Content, Creators' Authenticity and 'Messiness' Are in High Demand. Link
• KO Insights / O'Neill, K. (2026). The Authenticity Premium: Why Consumers Are Rejecting AI-Generated Content. Link
• Newswise / New York Institute of Technology. (2025). 2026 Will Require Brands to Balance AI and Authenticity. Link
• Shi, Y., & Jiang, Z. (2026). Consumer Responses to AI Disclosure Labels: The Role of Novelty and Authenticity. SAGE Journals. Link
• ResearchGate. (2026). An Analytical Study on the Impact of AI-Generated Content on Brand Authenticity and Consumer Trust. Link

‍