Jason Wolfe warns that AI safety needs broader coordination
An OpenAI alignment practitioner urges coordinated caution while welcoming his employer’s recent actions.
this is not a problem that can be solved by any one company (or country) in isolation.
About these dates
Statement date displayed on X in the inspected browser; not converted to UTC.
Jason Wolfe, whose OpenAI work includes the Model Spec, responded to Hubinger’s warning without assigning extinction odds. He described several harmful trajectories and feared the pace would make avoiding them difficult. This is Wolfe’s personal position, not an OpenAI policy announcement. Jason Wolfe ↗OpenAI ↗
He welcomed recent OpenAI actions while arguing that safety required coordination beyond one company or country. His reply favored coordinated caution over unrestricted development, rather than an unconditional shutdown demand. Jason Wolfe ↗Jason Wolfe ↗
Sources & attribution
- Commentary 8 Sept 2026Personal warning about development and coordination ↗
Jason Wolfe. Original source inspected. Personal views.
- Commentary 9 Sept 2026Reply on coordinated caution ↗
Jason Wolfe. Original source inspected. Personal views.
- First-party report 25 Mar 2026Inside our approach to the Model Spec ↗
OpenAI. Official article credits Jason Wolfe as its author.