A fabricated medical condition, bixonimania, was invented by a researcher and uploaded as two fake papers to an academic server. The condition was described as an eye disease that does not exist.
The researcher included absurd acknowledgments and a clear statement indicating the work was fictional. These red flags were designed to be obvious. Despite these warnings, major AI systems treated bixonimania as real for several weeks.
One chatbot claimed the disease was caused by blue light. Another reported a specific prevalence rate that was entirely fabricated.
A third advised users on matching symptoms, as if the condition appeared in medical textbooks. None of the systems detected the hoax. The fake study later appeared in a peer-reviewed journal.
The journal retracted the issue after intervention, but the paper had already spread. It had passed through both human reviewers and AI systems without anyone verifying whether bixonimania was real.
This incident is not merely about one researcher’s prank. It highlights a broader trend where AI-generated references are cited without verification. People are accepting what machines produce and acting on it.
This trend becomes more dangerous as AI moves into sensitive areas such as drug evaluation, patient consultation, and medical decision-making. The researcher’s experiment functioned as a vulnerability test.
It exposed a system that trusts and repeats information without checking. AI systems fell for the hoax because they are designed to generate plausible-sounding content, not to verify facts. They do not know what is true; they only know what appears true.
Human reviewers also failed. The fake paper made it into a peer-reviewed journal, meaning actual people read or skimmed it and deemed it publishable.
The retraction occurred only after someone intervened, by which time the paper had already circulated. The episode has sparked focus on verification, but focus is not the same as action. The same systems fooled by bixonimania remain in use in academic research, hospitals, and government agencies. The same human reviewers who missed the hoax are still reviewing papers.
No new rules or checks have been implemented. The problem is not that AI systems are imperfect.
That is widely known. The problem is that people treat their output as fact. A doctor asking a chatbot about a disease might receive a confident answer about a nonexistent condition.
A researcher seeking references might get citations to papers that were never written. A patient searching for symptoms might be told they have bixonimania.
The need for verification processes has been highlighted, but highlighting a need is not the same as meeting it. The systems that spread the fake disease are still online. The servers hosting fake papers still accept uploads.
The journals publishing without checking continue to publish. Bixonimania does not exist, but the vulnerability it exposed is real and remains unfixed.
{“@context”: “https://schema.org”, “@type”: “NewsArticle”, “headline”: “Fake Eye Disease Hoax Dupes Major AI Systems for Weeks”, “description”: “A fabricated medical condition, bixonimania, was invented by a researcher and uploaded as two fake papers to an academic server. The condition was described as “, “datePublished”: “2026-06-26T09:30:00+00:00”, “author”: {“@type”: “Person”, “name”: “Ariel Antonio”}, “publisher”: {“@type”: “Organization”, “name”: “InfoPulse Today”, “url”: “https://newsfeed.infopulsetoday.com/”}, “image”: “https://americansuntimes.com/wp-content/uploads/sites/19/2026/06/companion_23047_2.jpg”}
