A user on Mastodon shared their experience testing Google's Gemini AI by asking it to identify individuals in a photograph. The AI reportedly fabricated details, including misidentifying a person as Infantino de Messi and creating a false narrative about Angela Merkel's companions at an event. This highlights concerns about the reliability and potential for hallucination in current AI models. AI
IMPACT Highlights potential for AI hallucination and unreliability in user-facing applications.
RANK_REASON User-generated content on a social media platform detailing a negative experience with an AI model, without originating from the AI provider or a primary news source.
Read on Mastodon — mastodon.social →
AI-generated summary · Google Gemini · from 1 sources. How we write summaries →