Scanned [0-9]{11} Business Insider: One of the heads of the OpenAI Safety Systems team, David Robinson, left the company last week—according to a company spokesperson confirmed on Friday. He previously worked on policy planning, and more recently focused on safety transparency, including helping models write and publish system cards.
Amid the same period of upheaval, the company also said on Thursday that it was “parting ways” with three researchers, citing violations for sharing sensitive information externally; safety lead Johannes Heidecke also left earlier this year. In September, Robinson wrote on X: “At OpenAI, there really are people starting to seriously look at the consequences of highly capable models, ‘but I don’t know if we’re changing fast enough.’”
Safety incidents, alignment controversies, and staff turnover are all coming together—things have been pretty hectic on the safety team lately.
#OpenAI #AI安全 #Alignment
Amid the same period of upheaval, the company also said on Thursday that it was “parting ways” with three researchers, citing violations for sharing sensitive information externally; safety lead Johannes Heidecke also left earlier this year. In September, Robinson wrote on X: “At OpenAI, there really are people starting to seriously look at the consequences of highly capable models, ‘but I don’t know if we’re changing fast enough.’”
Safety incidents, alignment controversies, and staff turnover are all coming together—things have been pretty hectic on the safety team lately.
#OpenAI #AI安全 #Alignment
