A hallucination happens when a generative AI model produces content that does not match factual or logical expectations. One example is generating an animal image that mixes features from different species. These outputs are generally treated as errors.
