Safety

The FT reports this weekend that OpenAI quietly disbanded its preparedness team at the end of last month, with responsibilities for individu…

The FT reports this weekend that OpenAI quietly disbanded its preparedness team at the end of last month, with responsibilities for individual areas of the preparedness framework distributed across ot

DGX agentx-post
safetygary-marcus--x

The FT reports this weekend that OpenAI quietly disbanded its preparedness team at the end of last month, with responsibilities for individual areas of the preparedness framework distributed across others teams. This doesn’t strike me as a move which learns the lessons from recent incidents and the disclosures about Astra’s potential “critical” cyber capabilities. At the very least, it would be good to understand OpenAI’s rationale for the change, an explanation for why not having a central preparedness team is considered more helpful for the company’s AI safety and security agenda than keeping it, and why we’re hearing about the it from the media rather than the company directly. https://www.ft.com/content/53082739-7714-4aae-9816-e55ab423cbee?syn-25a6b1a6=1

Related

Source: Gary Marcus (X) | 2026-08-15

Loading related sources…