Worse still, some slop has become practically indistinguishable from reality, making it difficult to trust anything online. People are navigating the internet as weary detectives, constantly trying to decipher what is real and what is fake. Over half (53%) of consumers distrust AI-generated search results and summaries, a 2025 survey by Gartner found. And most people (70%) are uncomfortable with AI-generated media, a separate global survey by the management consultancy Baringo reported.  

The new battleground 

This erosion of trust has serious consequences for Europe’s major advertisers, retailers, media groups, and technology companies, many of which rely heavily on digital channels to reach consumers.  

It’s cast a paranoid pall over what are supposed to be engaging online experiences, explains Publicis Groupe’s Connected Media CEO, Niel Bornman. “A significant portion of people, particularly the younger generation, now operate under the assumption that everything they see online is fake,” he says. “This skepticism makes it harder for brands to establish genuine connections. It’s also a lot more expensive to get people’s attention.” 

According to Bornman, fake reviews and AI-powered search engines have become “the newest battleground” for brands competing online. “Some businesses have seen organic search traffic decline by between 5% and 35% as AI answer engines provide instant responses that stop users from visiting official websites,” he says. 

In response, brands are being pushed to spend more on pay-per-click advertising while simultaneously feeling pressure to use AI themselves to produce content at the scale necessary to “feed the machine” and remain visible in search rankings. “It’s a difficult dilemma,” Bornman continues. “Brands desperately want to avoid an AI-slop scandal, yet they also need to exist within these systems, appear as answers to consumer questions, and capture attention in an increasingly competitive environment.” 

Businesses are walking a fine line here. LinkedIn recently announced a crackdown on “generic” content that “lacks authenticity and originality,” even as it rolled out a suite of generative AI features, including a “rewrite with AI” button embedded directly into its post composer. 

The authenticity crisis 

One industry that is keenly aware of these problems is publishing. In March, Hachette withdrew the novel Shy Girl following allegations that sections of it had been generated using AI. The author denied directly using the technology, claiming instead that an editor inserted machine-generated passages into an early draft. Semantics aside, the controversy exposed growing anxiety about the industry’s ability to identify AI-generated material in manuscripts. 

“It’s the Wild West,” says Dan Conway, chief executive of the Publishers Association. “Large language models are hoovering up everybody’s content and using it with reckless abandon. The second a Premier League football club signs a major player, dozens of AI-generated biographies suddenly appear on Amazon. That may sound relatively harmless, but when the same inaccuracies appear in medical or educational material, the consequences become far more serious.” 

Senior leaders across industries are scrambling to contain the damage. Substack’s Chris Best warned of a coming “slop future,” while YouTube’s chief executive, Neal Mohan has publicly identified “managing AI slop” as a major priority for the platform in 2026.  

Yet there is little consensus about what should actually be done about it. 

How to mop up the slop 

“One way to regulate AI slop is to make it harder to profit from. If it can’t be monetized, the incentive to produce it at such a relentless rate disappears,” says Conway. 

But a second solution, one that is rapidly evolving into a new market of its own, is technology designed to distinguish authentic work from machine-generated content.  

Companies are investing heavily in AI-detection tools and verification systems capable of tracing where online material originates. Pinterest has introduced labels for AI-generated images, while Spotify has reportedly removed millions of bot-generated tracks using spam-detection systems. 

“There’s been an explosion in the use of AI-detection technologies,” says Bornman. “Executives are asking: what does this mean for my brand? Do I need to invest in it?” 

This may be only the beginning. From December 2026, the European Union’s AI Act will require many forms of AI-generated content to include digital watermarking—hidden digital markers indicating that the material was created using AI. More futuristic verification systems are also emerging, including blockchain-based provenance tools capable of maintaining verifiable records of how digital content was created and modified. 

Some tech leaders are comparing the rise of AI verification to the emergence of cybersecurity in the early 2000s. “Much like antivirus software was designed to stop malicious programs entering a PC, organizations are increasingly using detection tools to filter out AI-generated content,” says Mel Morris, Candy Crush and chief executive at Corpora.ai, an AI research engine.  

And yet, Morris warns, AI verification systems remain fallible. “Just as antivirus programs could miss countless threats while wrongly flagging harmless files, AI detection tools are often unreliable,” he says. 

“The answer to the machine may ultimately involve more machines,” Conway adds. “But it’s fair to say these tools won’t catch everything.” 

The dark side of detection 

The problem is that AI detection is far from an exact science. Most of these tools cannot definitively determine whether something is human- or machine-made; instead, they estimate the probability that content might be AI-generated. Inevitably, this creates false positives, where entirely human work can be incorrectly flagged. 

That problem can disproportionately affect people whose writing styles fall outside conventional patterns. Wipro’s global chief privacy officer, Ivana Bartoletti, is not a native English speaker. She says her writing tends to be very structured and concise, often relying on bullet points and short, direct sentences. When she runs her own work through AI detection systems, she says it is frequently flagged as machine-generated. 

“There is a serious discrimination issue at play,” she argues. “People who are neurodivergent, non-native English speakers, or simply write in a more formulaic way are far more likely to be incorrectly identified as using AI.” In hiring and corporate environments, this could lead to biased or unfair decisions, with algorithms penalizing people not because they used AI, but because their communication style happens to resemble it. 

Then there’s the thorny question of where acceptable AI assistance ends and slop begins. “What happens if something is 5% AI-generated?” she asks. “Is that acceptable or is it slop? What about 50%? It’s no longer a black-and-white issue.” 

Ironically, Bornman believes the systems designed to restore trust online may ultimately make the internet feel even more exhausting and suspicious. Instead of instinctively trusting what they see, users are increasingly being asked to verify authenticity for themselves, turning ordinary online activity into a constant exercise in doubt.  

Read more: The death of the billboard: Amsterdam’s ad crackdown is part of a much bigger European shift

“AI detection and verification make consumers work harder to build relationships with brands, in what was previously an effortless experience,” he says.  

The unwinnable arms race 

Already, the outlines of an AI-detection arms race are beginning to emerge. Developers are building tools that deliberately inject human-like mistakes into AI writing or scrub away the linguistic patterns detectors associate with chatbots.  

“If there was a detection technology that worked, people would simply build better AI tools to fool it,” Morris says. “It’s a cat-and-mouse game where both the content being generated and the tools used to detect it are powered by AI.” 

In the corporate world, some executives are already being coached on how to phrase earnings reports and public statements carefully enough to avoid triggering AI-powered sentiment-analysis systems searching for signs of weakness or instability. 

Complicating matters further, Morris argues, is the way human writing itself is beginning to change. “People are consuming more AI content and unconsciously learning to write in its style,” he says. “That will make it progressively harder for detection systems to distinguish between a person imitating AI and genuine AI output.” 

“At some point, we have to move beyond the assumption that simply because something is AI-generated, it is automatically worthless”

Mel Morris, Candy Crush and chief executive at Corpora.ai

One of the greatest dangers in this escalating technological arms race is the risk that authentic human work will be falsely identified as synthetic. “Students have reportedly faced disciplinary action after genuine essays were flagged as AI-generated by unreliable detection software,” Morris says. “While writers and professionals increasingly find themselves having to prove their humanity to algorithms.”  

Fighting the wrong battle? 

For some, the obvious solution may be to suppress or ban AI-generated material altogether. Major platforms, including Google, are already moving to de-emphasize content they perceive as low-quality or mass-produced by AI systems. 

But Morris argues that such approaches risk becoming blunt instruments. “A binary gate is dangerous,” he says. “Plenty of legitimate businesses now use tools like Claude or Gemini to help build websites, draft copy or streamline workflows. The presence of AI does not automatically make information inaccurate or low-quality.” 

The focus therefore shouldn’t be on eliminating AI from the internet altogether, he argues, but learning how to live with synthetic content without turning the web into a place where nobody trusts anything, or anyone, anymore. “Instead of obsessing over whether content was created by a human or a machine, businesses should focus on whether the information itself is accurate and reliable,” Morris says. “At some point, we have to move beyond the assumption that simply because something is AI-generated, it is automatically worthless.” 

Bartoletti agrees that detection technology alone cannot solve the world’s slop problem, particularly as studies increasingly show that people struggle to distinguish AI-generated content from human-created work—and often find both equally credible. “We need a combination of technological safeguards and human-centered approaches, including education, regulation, and organizational protocols,” she says. “These systems are only as strong as the frameworks surrounding them.” 

Simon James, global VP of data science and AI at Publicis Sapient, argues that while a handful of companies may try to differentiate themselves by rejecting AI-generated content altogether, most organizations will continue integrating AI tools into their operations in some form.  

For James, the answer lies not in state regulation but in self-regulation, with brands deciding for themselves what values they wish to defend in a slop-filled digital economy.  

“By the time legislation, such as the European Union’s AI Act, is fully implemented the technology underpinning it may already have evolved beyond recognition,” he says. “I think one of the most important things is self-declaring what is generated by AI, as opposed to pretending it’s not, or leaving the consumer under the impression that it might not be.” 

What the internet becomes in this age of AI slop will depend largely on how companies choose to respond to it. “Maybe it’ll finally become more intelligent, efficient and creative,” says James. “Or maybe, as the conspiracy theory goes, it is already dead.”