Generative AI conversational systems like ChatGPT have limitations such as hidden biases, factual errors, and hallucinations. These issues can be induced by injecting prompts into the system. In a blog post by Rafe Brena, PhD, strategies to prevent these issues are discussed. The full blog post can be found on Medium.

source update: This is How to Stop ChatGPT, Bing, Poe, and You from Hallucinating – Towards AI