GPT-4o Updates Have Made AI Personality 'Too Sycophant-y And Annoying', Says Altman

OpenAI Chief Executive Officer, Sam Altman, has acknowledged that recent updates to the GPT-4o model have resulted in a personality that is "too sycophant-y and annoying", promising swift fixes.

In post on X, Altman expressed concerns about the latest updates to the GPT-4o model. "The last couple of GPT-4o updates have made the personality too sycophant-y and annoying (even though there are some very good parts of it), and we are working on fixes asap, some today and some this week. At some point will share our learnings from this, it's been interesting."

This comes shortly after Altman announced a significant update to GPT-4o on April 26, aimed at enhancing both its intelligence and personality. "We updated GPT-4o today! Improved both intelligence and personality," Altman posted on X.

The GPT-4o model is part of OpenAI's ongoing efforts to refine its AI systems. Alongside GPT-4o, OpenAI has recently released o3 and o4-mini, the latest models in its o-series designed to think more deeply before responding.

The updates have sparked discussions among users and experts. Derya Unutmaz, a Twitter user, suggested, "Ideally, ChatGPT could have an emotional temperature tuner that shifts the tone from sycophantic to unobsequious. Eventually, it’d also be nice to have other AI personas to choose/tune from like forthright analyst, critical reviewer, devil’s advocate nonchalant sidekick, etc."

Another user, Matthew Berman, queried, "Is the personality crafted at the system message level or at post training?"

ChatGPT is an advanced artificial intelligence language model developed by OpenAI.