84 Gems

82620 Points

Badges: Rookie

OpenAI’s Soul Gets a Shakeup

OpenAI merges key research team shaping AI personality into core model development.

OpenAI is folding its Model Behavior team into the Post Training group, bringing closer focus on how its AI models act and feel when talking to people. The team’s leader, Joanne Jang, is starting a new research unit called OAI Labs to explore fresh ways humans and AI can work together.

The Model Behavior team helped shape AI personality, tackle bias, and reduce sycophancy.

Now reporting to Max Schwarzer, the group will work directly with core AI improvements.

Jang’s new OAI Labs will design novel interfaces beyond standard chat interactions.

This shift shows OpenAI sees AI personality as central to its future technology, especially as it faces pressure from users and lawsuits over how its models respond in sensitive conversations.