OpenAI has published a postmortem on the recent sycophancy issues with the default AI model powering ChatGPT, GPT-4o.

OpenAI CEO Sam Altman said that the company would roll back an update that users complained made ChatGPT 'too sycophant-y.' ...

GPT-4o is not a new model—OpenAI released it almost a year ago, but the company occasionally releases revised versions of ...

Meta Platforms Inc. lanzó el martes una aplicación independiente de inteligencia artificial para competir con otros chatbots ...

OpenAI adds shopping features to ChatGPT, offering product recommendations with images, prices, and direct purchase links ...

Users have been complaining the GPT-4o has become so enthusiastic that it's verging on sycophantic. The change appears to be ...

Perplexity's CEO Aravind Srinivas used the early days of Elon Musk and SpaceX—when the billionaire refused to give up even ...

After OpenAI changed ChatGPT to be less constrained around sensitive topics, TechCrunch experimented to see what minors could ...

OpenAI rolled back its latest ChatGPT GPT-4o update after users reported overly sycophantic responses. The company is now ...

OpenAI has rolled back a recent update to its GPT-4o model used in ChatGPT, following widespread criticism from users who ...

Sycophancy in AI refers to a model's tendency to align with the user's views rather than sticking to factual accuracy ...

The decision follows days of backlash on Reddit, X, and elsewhere, where users had labeled ChatGPT's tone as disingenuous and manipulative.