Act todayWhat happenedResearchers found a high-severity flaw in ChatGPT that exposed private conversations to attackers.
Who is affectedPeople and businesses who use ChatGPT, especially those who share personal, client or company information in chats.
What to do- 1Avoid typing passwords, account numbers or client secrets into any AI chat.
- 2Review your chat history and delete conversations with sensitive details.
- 3Turn on two-step sign-in for your ChatGPT account.
- 4Check OpenAI's official notices for fix updates and follow any advice given.
Researchers have found a high-severity flaw in ChatGPT that exposed private conversations to attackers. GBHackers News reported it on November 2, 2026. In plain terms, chats that people expected to stay between them and the AI could have been seen by someone else.
What happened
ChatGPT is the AI chatbot from OpenAI. People use it to draft emails, answer questions and work through problems, often by pasting in real details. Researchers rated this flaw high severity, a label used for serious security problems. The concern is that the exposure reached beyond the user, to attackers.
Who it affects
Anyone who uses ChatGPT could be touched in theory. The people with the most to lose are those who share personal, client or company information in their chats. A business owner who pastes in a client contract is in a different position from someone asking for a dinner recipe.
What is not known yet
The available reporting is thin. It does not say how the flaw worked, how long it existed, or whether attackers actually used it. It also does not say how many people were affected or whether a fix is in place. Those details come from the researchers, and they may change as more is published.
What to do
The alert box above lists the steps. The short version: treat any AI chat as something that could be exposed, and keep sensitive details out of it.
What happens next
Watch for an official response from OpenAI and for more detail from the researchers. Any fix notice or advice from the company is worth reading when it appears.
The takeawayTreat every AI chat as something that could be exposed, and keep passwords, client data, and health or money details out of it.