The institute is currently organizing a multi-lab AI safety workshop for next year and is in discussions with other frontier AI labs and safety organizations to establish collaborations. Tsimerman described the need for action on AI safety as extremely urgent in his announcement on Oct. 15.
