OpenAI has temporarily shut down parts of the popular ChatGPT service due to a bug exposing users’ chats. The bug was found by a Reddit user and confirmed by OpenAI to Bloomberg.
On a Reddit post, user Competitive-Hair-311 asked if ChatGPT had been hacked as they saw conversations with ChatGPT that they said they never had. These conversation topics include Xi Jinping’s Six Principles, China’s Ideology, Chinese socialism development, and other topics written in Chinese.
To Bloomberg, an OpenAI spokesperson said, “the titles were visible in the user-history sidebar that typically appears on the left side of the ChatGPT webpage. The chatbot was temporarily disabled after the company heard these reports. The substance of the other users’ conversations was not visible.
If you open ChatGPT’s chat, the Ai won’t be able to reply with an answer as it’s currently down. In addition, the company explains that History is temporarily unavailable as “we’re working to restore this feature as soon as possible.”
OpenAi has made the news this year due to the wonders of AI chat available with ChatGPT. More recently, the company unveiled GPT-4. A key difference between GPT-4 and GPT-3.5 (the language model that powers ChatGPT) is that GPT-4 is capable of accepting image inputs in addition to text. For example, if you show GPT-4 an image of eggs, flour, and milk and ask it what you can make with those ingredients, the language model will “see” the image and list several potential recipes.
GPT-4 is also significantly smarter than previous models. According to OpenAI’s research paper, GPT-4 passed a simulated bar exam with a score similar to the top 10% of test takers. ChatGPT scored in the bottom 10%. The multimodal language model also scored in the 99th percentile of the Biology Olympiad compared to the 31st percentile for ChatGPT.
With ChatGPT currently down, BGR will let you know once the chat and History features are back to users.
