ChatGPTthe language model for artificial intelligence GPT-3 from OpenAIreported its first security breach on March 22nd, and although Sam Altman -CEO of the company- reported this problem immediately without providing too many details, now the company has been responsible for providing more data regarding the scope of the vulnerability and the consequences it would have had for its users.
In the beginning, OpenAI noted that the breach allowed a small group of people to view the conversation history titles of others, but really this flaw in the system would have also exposed personal data of those users, the most sensitive being the banking information of those who contribute to the payment plan to get more benefits from the platform.
ChatGPT breach was more serious than previously thought.
OpenAI mentioned on March 20 that it had momentarily shut down the servers of ChatGPT in order to fix a bug that existed in the open source library, which allowed some users to read the history titles of others. However, the firm has now provided more details about all the data that was exposed, which turned out to be much more than initially thought.
“In the hours before closing ChatGPT on Monday, it was possible for some users to see the first and last name, email address, payment data and the last four digits (only) of the card number, as well as the expiration date of other users.”he said. OpenAI in its press release, clarifying that “at no time have the full card numbers been exposed”.
The time frame the company notes in its statement refers to the nine hours in which user information was exposed.
OpenAI responded quickly to the ChatGPT crisis.
According to the company’s research, the security breach of ChatGPT Would have affected 1.2% of payment plan users. Plus with the disclosure of their personal data to other users. These affected individuals have already been contacted by OpenAI. However, the company points out that it is highly unlikely that other users would have seen the private information since they would have had to have performed such specific activities as opening the confirmation email. ChatGPT Plus or log in to “Manage My Subscription” on the platform between 1:00 a.m. and 10:00 a.m. Pacific time slot on Monday, March 20.
OpenAI assures that the security breach has now been fixed and. ChatGPT is now fully operational again, with chat history restored and no problems. The company says it verified its solution to the bug, added checks to the user data request and improved the robustness of its infrastructure for better platform security.
We recommend METADATARPP’s technology podcast. News, analysis, reviews, recommendations and everything you need to know about the technological world.