← Back to live feed1 story
Micah Carroll, a leader in OpenAI's RSI/misalignment Preparedness sub-team, said the group is active and recruiting for research into monitoring and loss-of-control risks. Carroll described reporting that the unit was dissolved in July as "incredibly misleading" and stated the sub-team has "never been more empowered" to conduct its work, despite lacking a designated head of preparedness.
The dispute over the unit's status follows reports from the Financial Times that OpenAI had redistributed the team's responsibilities. These internal structural disagreements emerge as OpenAI manages a series of senior departures—including GPU systems engineer Scott Gray and revenue chief Denise Dresser—while preparing for an initial public offering.
Sign in to suggest edits
Key sources
- SOURCE@jachiam0“These are incredibly misleading headlines – @OpenAI Preparedness is very much alive and well by any meaningful definitio…”x.com
- SUPPORT@polynoamial“we are incredibly bandwidth bottlenecked and there is a lot of support for almost any impactful misalignment work you can think of”x.com
- SUPPORT@zeffmax“I reported last week that openai currently doesn't have a head of preparedness, but the subteams continue to operate and report to its safety advisory group”x.com
- SUPPORT@_nathancalvin“the level of churn we keep seeing in the structure/personnel status of safety work at OpenAI (and the incidents themselves) speaks to something real and concerning”x.com
- SUPPORT@kaicathyc“His team is doing critical research on monitoring and loss-of-control risks in an era of rapidly advancing capabilities”x.com