Businesses are battling something like a pervasive digital epidemic. AI slop is a broad term that refers to the endless flood of shoddy material generated by algorithms that has spread to almost every corner of the internet. Publishers are facing a flood of copycat books and fabricated reviews. The trusted online resources for everyday answers are flooded with questionable AI wisdom. And as if that wasn’t enough, fake bands are infiltrating your playlist.
To make matters worse, some slants are virtually indistinguishable from reality, making it difficult to trust anything online. People navigate the internet like tired detectives, constantly trying to decipher what’s real and what’s fake. According to a 2025 study by Gartner, more than half (53%) of consumers do not trust AI-generated search results or summaries. Another global survey by management consultancy Baringo found that most people (70%) are uncomfortable with AI-generated media.
new battlefield
This decline in trust is having a severe impact on Europe’s major advertisers, retailers, media groups and technology companies, many of which rely heavily on digital channels to reach consumers.
It gives off a paranoid coldness to what should be an engaging online experience, said Neil Bornman, CEO of Connected Media at Publicis Groupe. “A significant portion of the population, especially the younger generation, are now operating under the assumption that everything they see online is fake,” he says. “This skepticism makes it harder for brands to make real connections, and it’s much more expensive to get people’s attention.”
Bornman said fake reviews and AI-powered search engines have become “the latest battleground” for brands competing online. “Some companies have seen a 5% to 35% drop in organic search traffic as the AI answer engine responds instantly and prevents users from accessing the official website,” he says.
In response, brands are under pressure to increase their spend on pay-per-click advertising, while at the same time using AI itself to “feed the machine” and produce content at the scale necessary to maintain visibility in search rankings. “It’s a difficult dilemma,” Bornman continues. “Brands are keen to avoid scandals over AI issues, but they also need to exist within these systems, show up as answers to consumer questions, and stand out in an increasingly competitive environment.”
Companies are walking a fine line here. LinkedIn recently announced a crackdown on “generic” content that “lacks credibility and originality,” even as it rolls out a series of generative AI features, including a “rewrite with AI” button embedded directly in the post composer.
Thauthenticity crisis
One industry that is acutely aware of these issues is the publishing industry. In March, Hachette withdrew the novel. shy girl Following allegations that some of it was generated using AI. The authors denied using the technology directly, claiming instead that editors inserted machine-generated text into early drafts. Semantics aside, the controversy has revealed growing concerns about the industry’s ability to identify AI-generated material in manuscripts.
“This is the Wild West,” said Dan Conway, chief executive of the Publishers Association. “Large language models seek out everyone’s content and use it recklessly. The moment a Premier League football club signs a key player, dozens of AI-generated biographies suddenly appear on Amazon. That may sound relatively innocuous, but when the same inaccuracies appear in medical or educational materials, the impact is far more serious.”
Senior industry leaders are working hard to contain the damage. Substack’s Chris Best warned of a “sloppy future” ahead, and YouTube CEO Neal Mohan publicly stated “managing AI slop” as a key priority for the platform in 2026.
But there’s little consensus on what to actually do about it.
How to sweep a slope
“One way to regulate AI slop is to make it harder to profit from it. Without it, there is no incentive to produce AI at such a relentless pace,” Conway says.
But the second solution, which is rapidly evolving into a new market of its own, is technology designed to differentiate between authentic works and machine-generated content.
Companies are investing heavily in AI detection tools and verification systems that can track the origin of online material. Pinterest introduced labels for AI-generated images, and Spotify reportedly used its spam detection system to remove millions of tracks generated by bots.
“The use of AI detection technology is exploding,” says Bornman. “Business owners are asking, what does this mean for my brand and do I need to invest in it?”
This may be just the beginning. From December 2026, the European Union’s AI law will require various forms of AI-generated content to include digital watermarks (hidden digital markers that indicate the material was created using AI). More futuristic verification systems are also emerging, such as blockchain-based provenance tools that can maintain a verifiable record of how digital content was created and modified.
Some technology leaders liken the rise of AI verification to the emergence of cybersecurity in the early 2000s. “In the same way that antivirus software is designed to stop malicious programs from entering PCs, organizations are increasingly using detection tools to filter AI-generated content,” says Mel Morris. candy crush and CEO of AI research engine Corpora.ai.
Still, Morris warns that AI verification systems are still prone to error. “Just as antivirus programs can falsely flag harmless files while missing countless threats, AI detection tools are often unreliable,” he says.
“The answer to machines may ultimately involve more machines,” Conway adds. “But it’s fair to say that these tools can’t capture everything.”
The dark side of detection
The problem is that AI detection is far from an exact science. Most of these tools cannot clearly determine whether something was made by a human or a machine. Instead, it estimates the probability that the content is included. maybe Generated by AI. This will inevitably lead to false positives and potentially completely human work being flagged incorrectly.
This problem can disproportionately affect people who write outside of traditional patterns. Ivana Bartoletti, Wipro’s global chief privacy officer, 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. She says that when she runs her work through AI detection systems, it is often flagged as machine-generated.
“There is a serious problem of discrimination,” she insists. “Neurodivergent people, non-native English speakers, or people who simply write in a formulaic way are much more likely to be mistakenly identified as using AI.” This can lead to biased or unfair decisions in recruitment and corporate settings, where algorithms penalize people not because they use AI, but because their communication style happens to be similar to AI.
There’s also the thorny question of where acceptable AI assistance ends and the downward slope begins. “What if 5% of something was generated by AI?” she asks. “Is that acceptable or lenient? How about 50%? It’s no longer a black and white question.”
Ironically, Bornman believes that the systems designed to restore trust online could end up making the internet even more exhausted and suspicious. Rather than instinctively trusting what they see, users are increasingly being asked to verify the authenticity for themselves, making normal online activity a constant act of suspicion.
Read more: Death of billboards: Amsterdam’s advertising crackdown is part of larger European changes
“AI detection and validation has empowered consumers to engage more in building relationships with brands that were previously easy to do,” he says.
An arms race with no chance of winning
Already, we are beginning to see the contours of an arms race in AI detection. Developers are building tools to intentionally inject human-like mistakes into AI writing and remove language pattern detectors associated with chatbots.
“If we had detection technology that worked, people would just build better AI tools to fool it,” Morris says. “This is a cat-and-mouse game where both the content generated and the tools used to detect it are powered by AI.”
In the corporate world, some executives are already being coached carefully enough on how to word their earnings reports and public statements to avoid AI-powered sentiment analysis systems looking for signs of weakness or instability.
Further complicating matters, Morris argues, human writing itself is beginning to change. “People are consuming more AI content and subconsciously learning to write in that style,” he says. “This will make it increasingly difficult for detection systems to distinguish between someone imitating an AI and genuine AI output.”
“At some point, we have to move beyond the assumption that just because something is generated by AI, it automatically becomes worthless.”
mel morris candy crush and CEO of Corpora.ai
One of the greatest dangers in this escalating technological arms race is the risk that real human work will be mistakenly perceived as synthetic. “Students reportedly faced disciplinary action after authentic essays were flagged as AI-generated by unreliable detection software,” Morris said. “Meanwhile, writers and professionals increasingly have to prove their humanity to algorithms.”
Are you fighting the wrong battle?
For some, the obvious solution may be to suppress or ban AI-generated material altogether. Major platforms, including Google, have already moved to de-emphasize content deemed low-quality or mass-produced by AI systems.
But Morris argues that such an approach risks becoming a blunt weapon. “Binary gates are dangerous,” he says. “Many legitimate businesses are now using tools like Claude and Gemini to help them build websites, draft copy, and streamline workflows. The presence of AI does not automatically mean information is inaccurate or of low quality.”
The focus, he argues, should therefore not be on eliminating AI from the internet completely, but on learning how to coexist with synthetic content without turning the web into a place where no one or anyone can trust it. “Instead of focusing on whether the content was created by humans or machines, companies should focus on whether the information itself is accurate and trustworthy,” says Morris. “At some point, we have to move beyond the assumption that just because something is generated by AI, it automatically has no value.”
Bartoletti agrees that detection technology alone won’t solve the world’s slop problem, especially as studies increasingly show that people have a hard time distinguishing between AI-generated content and human-authored works, and believes that in many cases both are equally reliable. “We need a combination of technical safeguards and a human-centered approach, including education, regulation, and organizational protocols,” she says. “These systems are only as strong as the framework that surrounds them.”
Simon James, global VP of data science and AI at Publicis Sapient, argues that while a few companies may differentiate themselves by rejecting AI-generated content altogether, most organizations will continue to integrate some form of AI tools into their operations.
For James, the answer lies not in state regulation but in self-regulation, where brands decide for themselves what values they want to uphold in a chaotic digital economy.
“By the time legislation such as the European Union’s AI law is fully implemented, the technology supporting it may have already evolved beyond recognition,” he says. “I think one of the most important things is to self-declare what is generated by AI, rather than pretending it isn’t or giving consumers the impression that it might not be.”
What the internet will become in this age of AI will largely depend on how companies choose to respond to it. “It may end up being more intelligent, more efficient, and more creative,” James says. “Or maybe, as conspiracy theories say, it’s already over.”
