In 1998, researchers Benway and Lane observed users searching a website for a specific, important piece of information. The exact link they needed was placed inside a large, bold, colorful banner at the top of the page, yet very few users noticed it. Users had learned to look straight past anything resembling an advertisement, a reflex that quickly became known as banner blindness (Don Norman, “Banner Blindness, Human Cognition and Web Design,” March 1999). Today, the digital landscape is exhibiting the early stages of a modern equivalent to this phenomenon: AI blindness.
While no phenomenon called AI blindness has been directly measured yet, recent studies read side by side suggest users are developing a kind of immunity to content they believe was made by AI. This structural shift reached a cultural milestone when Merriam-Webster named ‘slop’, defined as low-quality digital content produced, usually in quantity, by AI, its Word of the Year in December 2025.
Suspicion alone is enough
The mere thought that a piece of content is synthetic triggers immediate audience rejection. According to an August 2025 Raptive survey of 3,000 nationally representative US adults, when people suspect content is AI-generated, they view it as 48% less trustworthy and 57% less authentic. Furthermore, they report a 60% weaker emotional connection to the material. This effect holds entirely regardless of whether AI actually produced the content; suspicion alone is enough to trigger the immune response.
Adding transparency labels does not solve this trust deficit. Research from the Nuremberg Institute for Market Decisions (NIM) surveyed 3,000 people across the US, UK, and Germany, finding that only 25% of participants think they can recognize AI-generated content. Broadly, only 21% trust AI companies, and a mere 20% trust AI itself. When advertisements were explicitly shown with an AI label, they were judged more negatively, and fewer people wanted to click through. Researchers termed this effect “small but meaningful.”
The attention gap and the closed circuit
A defining feature of low-quality AI content is “the gap between the attention it captures and the attention it sustains,” according to the Columbia SIPA Institute of Global Politics (June 2026). Camille François, a co-author of the Columbia report, warns that social platforms risk becoming closed circuits where creators and algorithms feed off one another, “while real audiences tune out.”













