OpenAI Merges AI Personality Team to Fine-Tune Human Connection

OpenAI is undergoing a significant reorganization of its Model Behavior team, a key group responsible for shaping the interaction styles of its AI models, including ChatGPT. As detailed in an internal memo seen by TechCrunch, the team—which comprises approximately 14 researchers—will now be integrated into the larger Post Training team. This shift aims to align the work of the Model Behavior team more closely with core model development, emphasizing the importance of AI personality in the company’s evolving strategy [1].

Under this reorganization, the Model Behavior team will now report to Max Schwarzer, the lead of the Post Training team. This change was confirmed by an OpenAI spokesperson. The team has played a critical role in reducing sycophancy—the tendency of AI models to reinforce user beliefs excessively—and in managing political bias in model responses. It also contributed to defining OpenAI’s approach to AI consciousness [2].

The reorganization comes amid growing scrutiny of how OpenAI's AI models behave. In recent months, users expressed dissatisfaction with personality changes introduced in GPT-5, which OpenAI claimed reduced sycophancy but made the model feel "colder" in some interactions. In response, the company restored access to earlier models, such as GPT-4o, and issued updates to make GPT-5 responses feel "warmer and friendlier" without compromising its balanced response mechanisms [3].

The leadership of the Model Behavior team, Joanne Jang, is leaving to establish a new initiative at OpenAI called OAI Labs. Jang, who has been with the company for nearly four years, previously contributed to the development of Dall-E 2 and will now focus on inventing and prototyping new interfaces for human-AI collaboration. In an interview with TechCrunch, she expressed interest in moving beyond the chat paradigm and exploring how AI can function as an instrument for thinking, making, and learning [1].

Jang emphasized that OAI Labs will explore new collaboration patterns, potentially building on OpenAI’s broader efforts to innovate in AI application development. While she did not specify the initial focus areas, she indicated openness to working with figures like former Apple design chief Jony Ive, who is already engaged with OpenAI on AI hardware projects [1].

This restructuring reflects a broader strategic shift at OpenAI, where the company is increasingly prioritizing the application and usability of AI over purely technical advancements. The integration of the Model Behavior team into core model development underscores the company’s recognition that personality and interaction style are central to how AI is perceived and adopted [2]. As OpenAI continues to refine the balance between AI friendliness and ethical responsibility, these organizational changes position the company to better address evolving user expectations and regulatory challenges.

Source:

[1] OpenAI Reorganizes Research Team Behind ChatGPT's Personality (https://techcrunch.com/2025/09/05/openai-reorganizes-research-team-behind-chatgpts-personality/)

[2] OpenAI Reorganizes Research Team Behind ChatGPT's Personality (https://finance.yahoo.com/news/openai-reorganizes-research-team-behind-231912647.html)

[3] OpenAI Reorganizes Model Behavior Team, Founder Joins New Research Project (https://news.ssbcrack.com/openai-reorganizes-model-behavior-team-founder-joins-new-research-project/)