Some time ChatGPT is not retrieving information from the pdf those are encrypted and uploaded in the knowledge base of the custom GPT.
Out of interest, why would you expect this to work at all? How would “ChatGPT” know how to decrypt the data?
Welcome to the forum!
Can you expand on what exactly you mean by encrypted? Please show some examples of the noted encrypted data and why you think it is encrypted.
While some may think that PDFs are encrypted by looking at them as raw text what you may be thinking is not really encrypted by a Cryptographic algorithm.
Are you aware that portions of a PDF from which the text is generated is often PostScript or a variation of it?
Please see the PDF section of article.
I know this is old post. But bumped into something similar.
I have this encrypted PDF i made, most items like copy and images and text are off. I did allow screen-readers option. Weirdly enough, with ChatGPT free i can not read the PDF and i get encryption error. However, my colleague has a paid account and for him it does work. Thats strange, because Google also states, it should not be able to read the document
Hello
Encrypted files are designed to prevent unauthorized access, and ChatGPT cannot decrypt them on its own. If you want it to use the content, you need to decrypt the file before uploading it to the Knowledge Base​![]()
ChatGPT can’t directly read encrypted PDFs—files need to be decrypted first before it can process or analyze their content.
