In a revealing episode of OpenAI’s official podcast, CEO Sam Altman urged users to approach ChatGPT with caution. Despite its widespread popularity, Altman emphasized that the AI chatbot is prone to “hallucinations”—a term used to describe confidently incorrect or misleading outputs.

“People have a very high degree of trust in ChatGPT, which is interesting, because AI hallucinates. It should be the tech that you don't trust that much,” Altman said.

He acknowledged that while ChatGPT continues to evolve with features like persistent memory and potential ad-supported models, it still has notable limitations. Altman also reversed his earlier stance on hardware, now suggesting that current computers weren’t designed for an AI-driven world and that new devices may be needed to support future systems.

Altman’s remarks come at a time when OpenAI is facing legal challenges from media outlets, including The New York Times, over content usage. He emphasized the importance of transparency and honesty about AI’s reliability, stating, “It’s not super reliable.”

The podcast also touched on the possibility of ads being integrated into ChatGPT. Altman said he’s not entirely opposed to the idea but stressed that any monetization must not interfere with the integrity of AI responses.

As AI becomes more integrated into daily life—from education and healthcare to entertainment and productivity—Altman’s message serves as a reminder that while AI can be powerful, it’s not infallible.