OpenAI's Response to the ChatGPT Data Leak
3 min read
OpenAI has confirmed that the recent data leak in ChatGPT was caused by a vulnerability in the open-source library, redis-py.
March 27, 2023 18:18
OpenAI has confirmed that the recent data leak in ChatGPT was caused by a vulnerability in the open-source library, redis-py. The leak occurred when some ChatGPT users were able to see what other users were using the AI chatbot for, as well as limited personal and billing information.
The issue started with an outage on March 20, but it turned out to be a more serious problem. During a nine-hour window on the same day, another ChatGPT user may have inadvertently seen your billing information when clicking on their own ‘Manage Subscription’ page. OpenAI notified 1.2% of ChatGPT Plus subscribers via email, but no evidence suggests that any customer information was viewed by more than one other ChatGPT user.
The internal investigation OpenAI conducted identified a bug in the Redis client open-source library redis-py. OpenAI uses Redis to cache user information in their server, Redis Cluster to distribute this load over multiple Redis instances, and the redis-py library to interface with Redis from their Python server. The bug caused many connections to return bad data after OpenAI made a change to their server on that Monday.
The bug has been patched, and OpenAI has added checks to make sure requesting users don’t get data belonging to other users. They have also improved the robustness and scale of their Redis cluster to reduce the likelihood of connection errors at extreme load. It has been estimated that ChatGPT has reached 100 million monthly active users, making this a wise course of action.
Consumers and businesses using ChatGPT should ensure it and OpenAI undergo the same third-party risk management process as any other application to avoid any future data leaks.