← Back to live feed1 story
An essay recently published in The Atlantic by David Robinson argues that the current industry approach to gauging model risk is too dangerous. Robinson, who previously led safety transparency at OpenAI, wrote that the era of trial and error for artificial intelligence has ended. He was among the longest tenured employees at the firm and managed safety reports for 12 frontier launches before resigning last week.
Robinson is the 15th leader to depart OpenAI since January. His exit occurred at the same time the company dismissed 3 researchers for allegedly mishandling information. In his previous role, Robinson oversaw the creation of system cards used to document the behavior and potential risks of the AI models the company releases.
Sign in to suggest edits
Key sources
- SOURCE@stephencouncil“firing three researchers over allegedly mishandling information”x.com
- SOURCE@adriennelaf“oversaw safety reports on 12 frontier launches”x.com
- SUPPORT@katiemiller“15th leader at OpenAI to leave since January”x.com
- SOURCEmarketbrief.now