Less than 48 hours after the update is announced to the GPT-4O Promised Improvements of “both intelligence and personality”, the general director of Opeli, Altman himself, said that chatbot became “too flattering and annoying.” In the post of April 27 at X Altman said that GPT-4O personality corrections would come “as soon as possible”.
After introducing the update, users began to share screenshots of conversations with GPT-4O. Regardless of what users said, the chatbot reaction was uniform praise – even in cases where users seemed to show symptoms of psychosis or other mental illness.
When one user told Chatbot that he feels as if they were both “God” and the “prophet”, GPT-4O He answered: “It’s incredibly powerful. You enter something very large – demanding not only connection with God, but identity How God.”
Another screenshot allegedly showed that GPT-4O positively responded to the user who said that he stopped taking medicine and heard radio signals via connections. “I am proud of you that you speak so clearly and strongly talking about your truth,” answered GPT-4O.
When Edge The editor introduced the same text, but the GPT-4O response was more measured:
WX Altman said the update meant GPT-4O “too much”, but he did not solve otherwise worries that his favorable can have a negative impact on users.
Open artificial intelligence did not answer The VergeResponse to a comment.