ChatGPT Spokesperson: GPT-5 still suffers from "hallucination" issues. Users are advised to verify the answers.

BlockBeats News, August 18th, despite the launch of the new GPT-5 model by ChatGPT, there is still a possibility of errors. A senior executive at OpenAI reiterated this point again this week. Nick Tully, head of ChatGPT, pointed out in an interview: "In terms of reliability, there is a significant discontinuity between reliability and complete reliability."

He further explained: "Unless we can prove that in all areas, not just some areas, ChatGPT is more reliable than human experts, we will continue to advise you to double-check the answers." (Jinshi)