Artificial Intelligence Bots Validate Fake Bixonimania Disease Created by Swedish Researchers in Misinformation Experiment

Researchers created the fake illness Bixonimania to test AI accuracy. Major chatbots and medical journals treated the fictional disease as a real condition.

By: AXL Media

Published: Apr 8, 2026, 4:36 AM EDT

Source: Information for this report was sourced from Nature

The Creation of a Digital Health Myth

Bixonimania began as a deliberate fabrication on March 15, 2024, appearing first in blog posts before being uploaded as fake preprints to academic networks. Almira Osmanovic Thunström, a medical researcher at the University of Gothenburg, developed the fictional illness to investigate whether artificial intelligence would treat unverified internet data as reputable medical fact. The experiment utilized a phoney lead researcher, Lazljiv Izgubljenovic, whose profile picture was generated by AI to add a layer of false professional credibility to the fictitious ocular condition.

Satirical Red Flags Ignored by Algorithms

Despite the scientific appearance of the papers, they contained overt clues intended to alert human readers to the hoax. The preprints credited funding to the Professor Sideshow Bob Foundation for advanced trickery and included acknowledgments to characters from the Starfleet Academy on the USS Enterprise. Furthermore, the documents explicitly stated that the entire paper was made up and that the test groups consisted of 50 non-existent individuals. These blatant indicators of fraud failed to prevent AI systems from indexing the content as valid medical literature within weeks of publication.

Automated Diagnoses From Leading Chatbots

The reach of the misinformation quickly extended to mainstream AI tools, which began advising users on the prevalence and treatment of the non-existent disease. In April 2024, Google’s Gemini suggested users visit an ophthalmologist for bixonimania, while Microsoft’s Copilot described it as an intriguing and rare condition. Perplexity AI even provided specific, hallucinated statistics, claiming the condition affected one in 90,000 individuals. These responses were generated both when users asked about the term directly and when they described general symptoms like eyelid hyperpigmentation.