A researcher created a fictional disease, complete with fabricated studies and obvious signs of forgery, yet AI models like ChatGPT, Gemini, and Perplexity took it seriously. The researcher even included a thank-you note to a professor from Starfleet Academy for a laboratory aboard the USS Enterprise, highlighting the AI's susceptibility to misinformation. AI
IMPACT Highlights the critical need for improved AI fact-checking and susceptibility to misinformation in generative models.
RANK_REASON The cluster discusses an experiment where AI models were fooled by fabricated research, serving as commentary on AI's susceptibility to misinformation.
Read on Mastodon — mastodon.social →
AI-generated summary · Google Gemini · from 2 sources. How we write summaries →