OpenAI Addresses Ex-employee’s Safety Claims

Jan Leike, head of alignment at OpenAI, resigned on May 17, citing irreconcilable differences with management

The CEO and president of ChatGPT manufacturer OpenAI, Greg Brockman and Sam Altman, respectively, recently took to X.com to discuss publicly the departure of Jan Leike, arguably the organization’s former chief safety officer. 

Leike was in charge of alignment. On May 17, he tendered his resignation, attributing it to divergent opinions with the organization’s leadership.

Lee’s allegations included that at OpenAI, “shiny products have taken precedence over safety culture and processes.”

Brockman and Altman promptly responded, visiting X.com within twenty-four hours of Leike’s post.

Brockman presented an extensive and convoluted discourse culminating in a tripartite approach to achieving safety alignment within the organization.

He commenced by expressing appreciation for Leike’s contributions to the organization before refuting assertions that OpenAI did not prioritize safety.

“First, we have raised awareness of the risks and opportunities of AGI,” Brockman wrote, pointing out that the company has “called for international governance of AGI before such calls were popular.”

As the concluding point of his lecture, Brockman stated, “The future will be more difficult than the past. “Our safety efforts must continue to advance in tandem with the stakes of each new model.”

Additionally, he suggested that the organization was not adhering to the central tech mantra of innovation and rapid progress. “As we progress in this direction, it is acceptable if our release timelines are delayed until we reach our release safety bar.”

Sam Altman, co-founder and chief executive officer, delivered a concise message but did indicate that he would elaborate further in the following days:

“He’s right.” About Leike’s remarks, Altman stated, “We are committed to completing the remaining tasks.”