There’s been a big shake-up over at OpenAI. Joanne Jang is departing her role with the Model Behavior team to take on the role of General Manager of OAI Labs. This new initiative will report to VTC’s executive director Mark Chen for the time being. Its purpose is to advance research and innovation in the development of AI interfaces. Jang’s move comes amidst increased scrutiny of OpenAI’s models. This announcement follows months of user outcry over personality changes in the most recent versions of their AI, ChatGPT and Bard.
Joanne Jang is the touchstone of OpenAI’s administration department. She played an enormous part in projects like DALL-E 2, the groundbreaking new image generation tool. Her insights into the Model Behavior team’s priorities have been instrumental in molding the personalities of OpenAI’s generative AI models, including ChatGPT. She has done amazing work to address the problem of sycophancy, where AI systems are overly agreeable to users, sometimes creating toxic user experiences.
For all OpenAI models starting with GPT-4, the Model Behavior team has been crucial in their iterative development. That didn’t stop them from inspiring the launch of GPT-4o, GPT-4.5 and GPT-5. At first, users decried the chillier demeanor of GPT-5 compared to its predecessor. This nuanced feedback reinforces the importance of Jang’s previous work, which centered on balancing personality traits in AI systems.
In her announcement on X, Jang expressed enthusiasm about her new project:
“i’m starting oai labs: a research-driven group focused on inventing and prototyping new interfaces for how people collaborate with ai.” – Joanne Jang
With OAI Labs, she wants to push things into more creative frameworks that go beyond chat or agent-based experiences. She stated,
“i’m excited to explore patterns that move us beyond chat or even agents — toward new paradigms and instruments for thinking, making,…” – Joanne Jang
Jang’s move to OAI Labs is a sign of OpenAI’s increasing dedication to innovating the future of human-AI collaboration technology. Her experience in addressing difficult challenges like sycophancy will no doubt serve her well in this exciting new position.
Following Jang’s transition, the Model Behavior team will now report to Max Schwarzer, OpenAI’s Post Training lead. Maker personalities and their avatars This amendment will accelerate the continued creation and improvement of AI avatars. It will continue to make them provocative while avoiding the trap of just playing back what users are saying.
Jang looks forward to all the new adventures that await her. She’s excited for the chance to work with Apple’s iconic designer Jony Ive, now creating a new family of AI hardware devices in partnership with OpenAI. This kind of collaboration has the potential to lead to out-of-this-world developments in user experience and interface design.
OpenAI has increasingly faced scrutiny over the behavior of its AI models, particularly regarding their interaction styles and user engagement. Users have provided important feedback in the chat about personality shifting and tone in models such as GPT–5. That kind of feedback points to the challenge of developing AI that really resonates with diverse audiences.