Read this AI slop

At IO+, we've always been transparent about the editorial team's use of AI. That's not just an ethical choice — since August 2, 2026, it's also a legal obligation under the EU AI Act. Yet that openness has a downside: the moment we mention AI, for instance by naming "Laio" as the author, the content gets dismissed as "AI slop" before anyone has even read the text. It doesn't matter that a human always reviews it. The stigma remains. This transparency paradox forces publishers into an impossible choice: comply with the law and risk rejection, or keep AI use quiet and preserve trust. Research and real-world examples show just how deep this dilemma runs, and why the solution lies not in rules, but in cultural change.

Law says: label AI. The public says: content has less value

Since August 2, 2026, transparency rules for AI-generated content have applied across the EU. Publishers must clearly indicate when AI has been used, especially for chatbots, deepfakes, and texts on matters of public interest that haven't had human review. The EU hopes this will restore trust, but reality proves stubborn. A study among Dutch consumers shows that 70% want transparency about AI use, yet that same transparency lowers purchase intent due to concerns about privacy and loss of control.

At IO+, naming the AI author "Laio" leads to rejection, even when the text has been checked by humans. The problem isn't quality, but framing: AI gets associated with "eeriness," an uneasy feeling that arises from the blurring of the line between human and machine. This effect is stronger for symbolic content, such as opinion pieces or analyses, where identity and emotion play a large role. The law mandates transparency, but the public responds with distrust — a paradox that calls for more than just labels.

Why we distrust AI, even when the text is accurate

Distrust of AI-generated content is deep-rooted and often irrational. A study by Science Says found that social media posts receive up to 8% fewer likes when labeled as AI-generated, even when the content is identical to human-made content. This stems from a combination of algorithm aversion — the tendency to trust algorithms less than humans, even when they perform better — and the "uncanny valley" effect, where something that is almost human becomes repellent precisely because of that.

It also plays a role that AI is seen as a shortcut: it suggests laziness or a lack of originality, whereas human effort is valued precisely because of the work involved. An incident at Ars Technica, where a journalist published fabricated quotes due to insufficient checking of ChatGPT output, reinforces this perception. Still, the solution isn't to hide AI. Research shows the stigma is weaker when AI is seen as a tool within a broader process, similar to Photoshop filters or translation software. The challenge is to normalize AI use without losing transparency.

The exceptions proving the rule

The AI Act distinguishes between different forms of AI use, with exceptions that reinforce the transparency paradox. For instance, AI-generated text doesn't need to be labeled if it has been reviewed by a human with editorial responsibility. This means IO+ technically doesn't need to use the "Laio" label, as long as the editor-in-chief has approved the text. Exceptions also apply to artistic, creative, or satirical content, and to standard edits such as spell-checking. Still, the practice remains complex.

The EU has developed four icons for labeling AI content, but usability tests show these are only effective when accompanied by a text label, such as "AI-generated." Moreover, responsibility remains with the publisher even when the icons are used. For publishers like IO+, this creates a bind: on one hand, you want to comply with the law and ethical standards; on the other, you want to avoid having your content rejected. The solution may lie in emphasizing human oversight.

Ultimately, the transparency paradox isn't a technical or legal problem, but a cultural one. Academic research shows that transparency about AI use leads to less trust in researchers, even when the AI was used only for supporting tasks. This calls for a cultural shift, in which AI is seen as a legitimate tool for specific tasks — just like a calculator or a word processor. In practice, this means publishers, editorial teams, and researchers themselves need to be transparent about their AI use, in order to break the stigma.

How media outlets are trying to break the paradox

Some media outlets are experimenting with ways to work around the transparency paradox. For example, Omroep Brabant used an AI clone of presenter Nina van den Broek to turn news articles into videos. The experiment was temporarily halted because older viewers found it confusing, while younger viewers actually appreciated it. DPG Media uses AI to generate headlines and edit text, but emphasizes that final responsibility remains with the editorial team. NOS accidentally used AI-generated images in its news broadcast, which led to criticism and the discontinuation of some AI tools. These examples show that transparency alone isn't enough: the users of AI — the editorial teams — also need to understand what they're using and what influence AI tools have. Only if you know how to handle a tool properly do you get good results from it. Ultimately, a "made with AI" label should be informative above all, not a warning.