OpenAI staffers flagged lax security before ChatGPT, five people say

The concerns described by the five people predate the public launch of ChatGPT in late 2022, meaning they were raised while OpenAI was still primarily a research lab rather than the consumer-facing company it became. According to those accounts, staffers took their worries not just to top leaders but to the board of directors, the body's highest oversight layer. Some employees came away from those exchanges concluding that management's attention was fixed on keeping pace with competitors rather than on tightening controls.

The specific gap workers pointed to was internal access: anybody at OpenAI who could reach its code could see the technical details underlying many company projects. The people said this made it relatively easy for outsiders, including foreign governments, to obtain sensitive information about the lab's technology.

When OpenAI disclosed in October that a recording device had been found in an executive conference room, a number of former employees reacted with recognition rather than shock, according to the account. That reaction ties the October incident back to concerns the people say were voiced years earlier.