The Authenticity Paradox: Why “Human-Made” Has Become the Most Powerful Claim in the Age of AI

The Authenticity Paradox

In 2025, Swedish fashion giant H&M announced that it would create “digital twins” of thirty of its models. These AI-generated copies would appear in its social media posts and marketing campaigns. H&M insisted that it was not replacing traditional photoshoots, and that the models would own the rights to their twins and be paid each time they were used. Critics still saw what the move made possible: a fashion shoot without actually having one. The photographer, the lighting crew and the location could all be replaced by a prompt. The backlash came quickly, from models, photographers and unions as well as from shoppers online, many of whom saw it as cutting corners, even though the images looked entirely professional.

The use of AI in social media ads and content has been growing since 2022, but it appears to be putting many consumers off. In one recent global survey, 31% of consumers said they trusted a brand less when they noticed AI-generated content in its marketing, compared with just 7% who trusted it more. It is much harder to determine which factors really drive consumer behaviour when people actually make a purchase. But it is no longer the obviously broken images and strange hands of early campaigns that are driving consumers away. It is the sense that something just feels off, as if brands no longer care enough to make an effort. If they can’t be bothered to advertise it properly, why should we buy it?

AI slop is low-quality, fake, rapidly produced AI content, and it is now everywhere. One study found that around one in five videos shown to a brand-new YouTube account were low-quality AI content. The appeal of AI to executives and content creators is obvious: faster production, cheaper visuals, smaller budgets and better targeting. But the shift is happening at precisely the moment that consumer tolerance is running out.

Research consistently shows that when identical images are labelled as either “human-created” or “AI-created”, people judge the supposedly human ones as more beautiful, more meaningful and more profound, even when they cannot tell the difference visually. What we value in creative work is not just the output, but the intention, effort and human experience we believe sit behind it.

For H&M’s critics, scrolling past a model who was never in the room felt like an uncomfortable contradiction for a brand in an industry that is built on human creativity.

Other brands are starting to pick up on this consumer sentiment. In 2025, Cadbury launched a completely AI-free campaign, using real chocolate on set with no CGI or AI. Global Brand Vice President Guilherme Ferreira said: “When it feels like everyone is turning to AI and CGI, we were proud to create the film using real life filming on set, utilising experts in their craft.” Feedback from both consumers and marketing analysts was very positive, but it remains to be seen whether this will become the norm. The difference in cost is still huge, and the effect on profits is still unclear.

On social media, resistance is growing louder. Théodore Cazals, a student from Paris, started an X account called “Insane AI Slop” after a fake Facebook image of bearded children beside a birthday cake in the middle of a busy road went viral with nearly one million likes. His account grew to more than 133,000 followers. Under viral AI videos, angry comments now often outperform the posts themselves. Because anger and outrage are among the biggest drivers of engagement, this pushes the content even further up the algorithm.

So will content continue down the route of AI, or will the backlash grow stronger?

In one scenario, the backlash shifts the market: AI-generated content starts to be perceived as low-end, and “human-made” as premium. The label could become as powerful as “organic” or “Made in Italy”: a sign of quality, care and trustworthiness that commands a much higher price.

A second, arguably more likely, scenario is that nothing fundamentally changes. Anger, as we have seen, feeds the algorithm just as effectively as admiration. Théodore Cazals’ 133,000 followers are, despite their best intentions, still scrolling. The platforms that profit from outrage have no incentive to reduce the content that provokes it, and many brands’ budgets will continue to favour the prompt over the photographer.

A third possibility is that people start turning away from mainstream platforms altogether. Trust is the cornerstone of social media. People scroll because they half-believe that what they see has some connection to reality. Once that belief erodes, the contract breaks. We may be approaching a point where AI slop drives users not simply towards anger, but towards smaller, human-moderated online communities, or away from screens entirely.

The fourth scenario may be the most unsettling of all. Researchers are asking not just whether people can tell real content from AI-generated content, but whether they will bother trying. Alessandro Galeazzi of the University of Padova warns that AI slop increases the “brain rot” effect: the gradual erosion of attention and critical thinking caused by constant exposure to low-quality content. In this scenario, nobody leaves, nobody fights back, and nobody particularly notices. The feed becomes background noise: occasionally stimulating, and entirely disconnected from anything that could be called reality.