Aye that’s a bad one, this is different from how paid API keys for applications to use ChatGPT in their backends were also hacked recently yeah? not stellar by them, sigh.

“Employees enter classified correspondences or use the bot to optimize proprietary code. Given that ChatGPT’s standard configuration retains all conversations, this could inadvertently offer a trove of sensitive intelligence to threat actors if they obtain account credentials.”

As I recall, the current advice from pretty much everyone is: Don’t give ChatGPT or any other LLM sensitive information ever. Even without stolen credentials, these things like to regurgitate information they have seen before. E.g. Grandma reading you windows keys to fall asleep.