By now, the challenges posed by generative AI are no secret. Models like OpenAI’s ChatGPT, Anthropic’s Claude and Meta’s Llama have been known to “hallucinate,” inventing potentially misleading responses, as well as divulge sensitive information, like copyrighted materials.
Q&A: Model disgorgement—the key to fixing AI bias and copyright infringement?
