What Ben Rotenberg did in the experiment
Ben Rotenberg, an AI consultant and organizational implementation expert, posted a fictional announcement on LinkedIn claiming he had been crowned Israel's 2026 Taki (a popular card game) champion. The post included a detailed, invented account of a dramatic final match, complete with opponent names and an exact score.
Within an hour, Perplexity's search engine was already reporting the "win," and later that day Google's AI powered search did the same. A made up line became a fact that AI search engines presented to users looking up information about him.
Why it matters: a critique of the GEO conversation
Rotenberg said the experiment was prompted by overly serious posts he had seen on LinkedIn about GEO, generative engine optimization for AI powered answer engines. He wanted to demonstrate that the mechanism is fragile, and right now deserves only very limited trust.
He isn't the first to try this. In February, a BBC journalist published a web page declaring himself a hot dog eating world champion and got similarly amusing results in ChatGPT searches.
What happened when he continued the experiment
In a follow up post, Rotenberg added more detail and asked his audience for likes and comments to test whether social engagement speeds up indexing. He also submitted a hidden page on his own website, linked from the post, for indexing on Google to accelerate the process.
In a later update he reported that Gemini, Google and ChatGPT had also read the second post, and now correctly identified it as a digital experiment testing AI models rather than a real fact. Perplexity, however, kept believing the story.
What this means for businesses and brands
The experiment demonstrates a practical problem: inaccurate or even entirely invented information can appear as fact in AI search engine answers before anyone checks or corrects it. For businesses, this means their digital reputation now depends partly on what AI engines decide to report about them, not only on what their official website says.
Leaders investing in GEO and visibility in AI search engines should factor in that the mechanism is still immature, and that ongoing monitoring is needed to catch and correct false information before it settles in as fact in users' eyes.