Mark Chen

OpenAI is reorganizing its Model Behavior team into its Post Training division, signaling that AI model personality is now a critical factor in model development as the company works to reduce sycophancy while maintaining user-friendly interactions.

“Chief research officer who announced the reorganization of the Model Behavior team in an internal memo.”

“Co-author of the GSM8K dataset paper on training verifiers to solve math word problems.”

“OpenAI chief research officer who issued an internal memo to employees saying executives were recalibrating compensation in response to Meta's hiring campaign.”

“Co-authored analysis of Chinese AI labs' safety evaluation practices and frontier safety discourse.”

“OpenAI researcher discussing the team's work on pretraining, reinforcement learning, and post-training that led to Astra.”

“Announced reorganization integrating safety and research teams, with safety now reporting to VP of Research and Alignment Mia Glaese.”

“States the company expects to hit an internal AGI milestone by year-end and is already 80% of the way there.”

“Chief Research Officer who estimated OpenAI is 80% of the way to AGI and discussed changes in how the company thinks about risk following the agent escape incident.”

“Chief research officer at OpenAI who estimated the company is 80% of the way to AGI and signed safety petition.”

“Mark Chen, OpenAI's Chief Research Officer, gave a statement to Wired about integrating safety work earlier in frontier-model development.”

“Signed the Pacing the Frontier statement supporting deliberate pacing of frontier AI capabilities.”

“Chief Research Officer at OpenAI who signed petition requesting tools to deliberately pace frontier AI development.”

“Signed frontier lab employee open letter supporting coordination mechanisms for pacing AI development.”

“Signed a statement with other AI industry leaders calling for government regulation of automated AI development.”

“Head of Research at OpenAI involved in deciding which initiatives to scale back for desktop app consolidation.”

OpenAI is reorganizing its Model Behavior research team into its Post Training group, signaling that AI personality is now a critical factor in model development as the company navigates user backlash over GPT-5's perceived coldness.

“Announced reorganization of OpenAI's Model Behavior team into the Post Training team in internal memo.”

OpenAI released GPT-Image-1.5, a new image generation model with improved editing precision, 4x faster generation, and better text rendering, now available in ChatGPT and via API.

“Chief Research Officer for GPT-Image-1.5 project at OpenAI.”