AI Chatbots and Humans Fooled by Fake Eye Disease 'Bixonimania'

Stockholm. The fake eye disease named 'Bixonimania', created to fool AI chatbots, has confused not only technology but also people. In 2024, a team of researchers led by Almira Osmanovic Thunstrom from the University of Gothenburg, Sweden, created this fictional disease to test how easily chatbots believe misinformation. Within weeks of the research team placing two fake studies about 'Bixonimania' on the internet, popular AI platforms like Microsoft's Copilot, Google's Gemini, and OpenAI's ChatGPT began presenting it as a real disease. Copilot called it a 'rare and interesting condition' while Gemini even claimed it was caused by excessive exposure to blue light. Even more surprisingly, some human researchers also cited this fake disease and its fictional studies as references in their research papers. The research team had left many clues to identify that this study was entirely fake. The author of the research paper was named Lazlalziv Izgoublenovik (which means liar in Slavic languages), and he was shown to be working at the non-existent Asteria Horizon University in California. Fictional names like Professor Sideshow Bob Foundation and Galactic Triad were used as the financial source for the study. Furthermore, the sentence 'This entire article is fabricated' was even written in the middle of the research paper. Despite such clear indications, the fact that AI and humans were fooled raises serious questions. Although the name of the eye disease, which includes the word mania related to mental health, is itself suspicious, chatbots directly included the data available on the internet in their answers without distinguishing between true or false. According to experts, this incident has exposed how dangerous it can be to blindly trust AI-based information and how weak the filtering capacity of scientific research is.