[Column] Michel van Velde: AI search is a sloppy journalist

AI driven search engines are transforming the information landscape faster than many organizations realize. Where we relied on Google for years, millions of people now turn to models such as ChatGPT and Google AI for direct answers. This creates a new dynamic that is often underestimated: AI search increasingly behaves like a journalist, but a sloppy one.

A good journalist selects sources, checks facts and builds a story that holds up. AI works differently. The model consults multiple sources, but without a journalistic conscience. It verifies nothing, does not assess the reliability of sources and composes an answer based on probability. Truth is no longer a checked and validated concept, but a construction shaped by visibility, repetition and semantic recognition. That creates risks and opportunities.

How AI search gathers information

Recent research by Nective Digital, published on Search Engine Land, shows that ChatGPT performs its own internet search in roughly one third of all cases. This means that AI does not only rely on the training data it was built on, but also actively retrieves new information.

It does this through longer and context rich search terms, often referred to as Fan Out Queries. These terms regularly contain words such as comparison, review and specific years such as 2025. AI looks for content that appears complete, current and semantically clear.

But these results are rarely verified. AI does not distinguish between reliable and unreliable sources. The journalistic layer of final editing is entirely absent. Various tests even revealed that some of the sources mentioned by AI never existed, yet the model presented them with full conviction.

Truth as the sum of visibility

Because AI does not verify but assembles information, a new form of influenceability emerges. The question is no longer simply whether the right information is online. The real question becomes whether that information is visible, consistent and semantically clear enough to be picked up by AI.

AI selects what it understands. Content that is clearly structured, uses consistent terminology and appears in multiple places is favored. Repetition feels like confirmation to a model. As a result, organizations that use inconsistent wording or publish different definitions risk having their story distorted. AI fills gaps with assumptions, combines old information with new elements and may associate a brand with themes that are completely incorrect. And because more users rely on AI answers, this confusion becomes visible in every interaction between humans and machines.

A new strategic playing field

The time when you wrote only for humans is over. You now also write for machines. AI search acts as a second reader, one without a sense of nuance but with a strong preference for patterns, structure, consistency and authority.

This demands a different approach to strategic thinking. Brands must decide which terminology they want to claim, how they articulate their positioning, which arguments they consistently repeat and how they build authority within the digital ecosystem. AI does not need to be convinced. It needs to be fed.

Brand consistency is no longer an aesthetic choice but a strategic requirement. A brand that tells its story differently in each context disappears from AI generated answers or is represented incorrectly. A brand that applies discipline to its narrative gains an advantage: AI recognizes the pattern and reproduces it.

Influencing AI without deception

Influencing AI search does not mean manipulating systems. It is about understanding and applying the rules of the new playing field. This starts with consistent terminology: what you call your product, your method and your claim. It then requires content that is logically structured, clearly written and semantically sound.

Authority is the third essential factor. Publications in professional media, whitepapers, case studies and consistent references build credibility. The more often a model encounters your brand in a serious context, the greater the chance that your information becomes part of the answers that AI generates.

You brief an AI system in the same way you would brief a journalist. The difference is that the journalist asks questions and checks facts. The AI does not. That is why your story must be correct from the start and consistent everywhere. This requires alignment between organization, process, content and technology.

Conclusion: if AI is a sloppy journalist, be the spokesperson who communicates clearly

AI search is a breakthrough, but at this stage still a journalist without an editor: fast, persuasive and at times dangerously inaccurate. Anyone who does not actively manage how their brand appears online leaves room for noise, errors and misinterpretations that may reach millions of people.

The only way to be visible and accurately represented in this new reality is to bring discipline, repetition and clarity to the story you tell. Brands that do this are recognized first by users, and increasingly by the AI systems that retell their story.

If AI is the sloppy journalist, you must be the spokesperson who provides the right story. Whoever controls the input, controls the output.

Michel van Velde  is CEO and Brand & AI Director at  Ablestar . He advises international brands on the intersection of brand strategy, digital innovation, and artificial intelligence.

 

 Follow Marketing Report on LinkedIn

 Subscribe to the Marketing Report Newsletter

 Send us your press releases, columns, interviews…

 Join Marketing Report in our LinkedIn Groups