OpenAI CEO Sam Altman has issued a strong public advisory, urging users to “not over-trust ChatGPT or any large language model”. In a recent conference appearance, Altman acknowledged that while ChatGPT is a powerful tool, it still occasionally "hallucinates"—producing incorrect or misleading information with high confidence.

“AI should be used with judgment. It’s a great assistant—but not a perfect source of truth,” said Altman.

He further raised concerns over current AI hardware bottlenecks, hinting that today's infrastructure may be struggling to keep up with the ever-growing demands of large model training and global inference loads.

Altman’s comments come at a time when ChatGPT usage is at an all-time high, and AI-generated content is becoming deeply embedded in education, media, and business workflows.

His advice underscores the importance of AI literacy and human oversight, especially as companies and individuals rely more heavily on generative AI for critical decision-making.