An AI hallucination is a situation in which a language model generates information that is factually incorrect, nonexistent, or fictional, yet is expressed in a confident and credible manner. This occurs because the model actually predicts the most likely continuation of the text based on learned patterns, rather than verified and current facts, and lacks a built-in mechanism to verify facts outside its training data. The risk of hallucinations can be reduced, for example, using the RAG technique, which forces the model to base its response on specific, verifiable sources.
AI Hallucination
The AI Act for Board Members: What Every C-Level Executive Needs to Know to Keep the Company from Paying Millions
6 min read
AI Governance in the Workplace: How to Set Rules for AI Use Before Regulators Do It for You
5 min read