“OpenAI Removes AI Safety Leader Aleksander Madry, a Onetime Ally of CEO Altman”, 2024-07-23 (; backlinks; similar):
[Altman response] OpenAI has reassigned the executive in charge of ensuring that its artificial intelligence won’t harm society to a different role within its research organization, the company told staffers last week.
The executive, Aleksander Madry, set up OpenAI’s Preparedness team last year to evaluate AI models for “catastrophic risks” before making them available to the public, according to the company. Last week OpenAI leadership told employees that Madry, who is also a professor at the Massachusetts Institute of Technology, would take on a “bigger role within the research organization” focusing on fundamental work in reasoning, but it doesn’t appear to be nearly as prominent as his last role, according to a person who saw OpenAI’s announcement.
…She [Lilian Weng] will report to Joaquin Quiñonero Candela, who previously spent nearly a decade at Facebook working on AI safety issues [background]…