@ai_and_law · Post #807 · 15.04.2026 г., 07:05
📖Fake Disease Experiment Exposes AI Misinformation Risks Researchers led by Almira Osmanovic Thunström at the University of Gothenburg created a fictional medical condition, “bixonimania,” and supported it with fabricated academic papers uploaded to a preprint server. The experiment aimed to test whether large language models would reproduce false information as credible medical knowledge. AI systems subsequently presented the non-existent disease as real, and the fake studies were later cited in peer-reviewed literature, including a journal published by Springer Nature. The case demonstrates how AI-generated or AI-amplified misinformation can propagate beyond initial sources into scientific and medical contexts. #AIRegulation#AIethics#Misinformation#AIsafety#DigitalHealth