Hi everyone,
I recently encountered an odd and concerning issue with ChatGPT that I’d like to share for feedback or investigation. I uploaded a PDF containing AI-related content and asked ChatGPT to make notes on it. Instead of summarizing the AI content, it returned detailed student information—including a name, father’s name, date of birth, enrollment number, course, university, and session—that wasn’t in the PDF at all. I don’t know the student, and this data seems completely unrelated to my input.
For comparison, I gave the same PDF and prompt to Grok (from xAI), and it correctly summarized the AI content without any extraneous details. This suggests ChatGPT might be introducing information it shouldn’t, which raises a privacy red flag for me. Could it be pulling sensitive data from somewhere else or hallucinating in a way that mimics real personal info?
Has anyone else seen this kind of behavior? I’m planning to report it via the Help Center too, but I’d love to hear thoughts from the community or OpenAI team on what might be happening here. It feels like a potential privacy issue that needs attention.