Home · Personal · Tech & Digital Life · Privacy & security

PARTLY

As of 13 August 2026, AI can only partly check a document for personal data before sending it.

This still needs a person who signs their name to it.

Can you do it?

5 minutesto a draft.

30 minutesto something you’d act on.

Cost, all in£0

Skill neededchat-fluent

Who has to check ityou

What the alternative costsThe supplied tool data gives no price for a specialist document-redaction service.

If this goes wrong, you send a document containing an overlooked identifier, hidden comment, image detail or piece of metadata.

What to actually do

  1. Hand it to a person

    The route this page recommends

    A person who owns the outcome does this end to end, worth it when the failure is dear.

  2. Use a tool built for this

    Second choice

  3. Do it yourself

    The distant third

    A chat interface, chat-fluent skill, and roughly 30 minutes until you can act on the result.

    How to actually do it

    1. Open the original document and make a separate working copy so the original remains unchanged.
    2. Check whether the file contains personal data that should not be uploaded, and remove unrelated pages or replace real values with placeholders where doing so will not defeat the check.
    3. Open a trusted chatbot, paste the prompt, and attach or paste the working copy.
    4. Compare every AI finding with the original document, including headers, footers, tables, signatures, images, handwritten marks and visible contact details.
    5. Open the document's properties, comments, tracked changes, hidden text, attachments and revision history, and remove anything that should not be sent.
    6. Use permanent redaction tools that delete or cover the underlying text, then close and reopen the redacted copy and search it for names, email addresses, numbers and other identifiers.
    7. Send the checked copy only after confirming the recipient, attachment and sharing permissions are correct.

    Prompt

    Check the document below for personal data before I send it. Do not reproduce the actual personal data in your answer. Instead, list each finding by page, paragraph, table cell, image or metadata field where possible, and describe its type, such as name, address, email address, phone number, account number, identification number, signature, location, date of birth or information about another person. Flag information that could identify someone when combined with other details. Separate definite findings from possible findings. Check the visible text, tables, headers, footers, images and document properties if they are available to you. Also tell me what you cannot inspect in this format. Suggest what to redact, but do not alter the document and do not claim that it is safe to send. I will compare your list with the original file before sending it.

Document:
[PASTE THE TEXT OR ATTACH A COPY OF THE DOCUMENT HERE]

    Open it prefilled in ChatGPT or Claude, or copy it into Gemini, which takes no prefill link.

What it gets wrong

  • AI cannot guarantee that it has found every identifier, especially in scans, handwriting, images, metadata or unusual file formats.
  • AI cannot know which information you are legally or contractually required to remove without the relevant context and rules.
  • AI cannot safely inspect a document without you deciding whether disclosing its contents to the chatbot is acceptable.
  • AI does not create reliable permanent redactions by merely suggesting text to hide or placing a black shape over it.

What caps this at PARTLY: private data access, verification cost and consent and privacy.

How we scored this

Five axes, each scored nought to two by hand: ten means AI carries the task cleanly, and the thresholds that turn a total into YES, PARTLY or NO are published in the methodology. Each axis name links to its definition.

AxisScore (0–2)
Output2
Inputs1
Verification1
Liability1
Effort delta2
Total7 / 10

The methodology and its thresholds are published in full.

FAQ

Can ChatGPT find personal information in a document?
Yes, it can identify many visible names, addresses, contact details and numbers in pasted text or an uploaded file. It can miss information in images, metadata, scans and unusual formatting, so compare its findings with the original.
Is it safe to upload a document with personal data to AI?
Not automatically. Check the service's privacy and retention settings, remove unnecessary pages or replace values where possible, and do not upload the document if disclosure is not acceptable.
Can AI remove personal data from a PDF?
AI can suggest what to remove, but a chatbot is not a guarantee of permanent redaction. Use a redaction tool that removes the underlying content, then reopen the result and check its text, images and metadata.
How do I check a document before sending it?
Scan the visible text with AI if appropriate, then check the original page by page for names, contact details, identification numbers, signatures, images and other people’s information. Also inspect comments, tracked changes, hidden text, attachments and document properties before sending the checked copy.

Nearby answers

All Privacy & security rulings

Assessed by gpt-5.6-luna (gpt-5.6-luna) on 2026-08-13, second-checked by an independent model. Wrong somewhere? Email [email protected] and it gets re-checked.

The newsletter

AI news, new answers and product picks, straight to your inbox.