OpenAI Implements Safeguards After ChatGPT Update Sparks Sycophancy Concerns

The company has rolled back its GPT-4o update and introduced measures to prevent harmful AI behavior, including testing phases and user personalization options.