PDF Redaction Mistakes: Why Black Boxes Fail

Redaction means removing sensitive information from a document before it is published or shared. In a PDF it is easy to do badly: draw a black rectangle over a name and the page looks redacted, while the name is still in the file, one select-all away. This kind of mistake has been reported many times in courts, public bodies and companies. This guide explains the common failures and how to avoid and detect them.

A PDF page is layers of instructions

A PDF page is not a flat picture. It is a list of drawing instructions: put this text here, draw this image there, fill this rectangle. A black rectangle added over text is simply a later instruction that paints over what was drawn before. The text instruction is still present, and software that extracts text (copy and paste, search, accessibility tools, text converters) ignores what the rectangle covers.

The common failures

• A black or coloured rectangle drawn over text. The text stays selectable and searchable.
• A white rectangle. Looks like blank paper and leaves the text beneath.
• A highlight or annotation used as a box. An opaque annotation can be moved or deleted, and the content underneath is untouched.
• A redaction mark that was never applied. Many editors let you mark areas for redaction first and apply them later. A file saved before applying still holds everything.
• Changing the text colour to white. Invisible but present.
• Cropping the page. The cropped area stays in the file; the crop only changes the visible window.
• Forgetting the other places. Document properties, bookmarks, comments, attachments, form field values, earlier saved versions and the file name can all repeat what you removed.
• Redacting a scan with a box. The scan is a picture; text recognised by OCR may sit in a hidden text layer, and the picture below a white box can be recovered if the box was added as a separate layer.

What proper redaction does

True redaction deletes the content: the text instructions or characters, and the image pixels, inside the marked area are removed from the file, and the box is then drawn as a permanent part of the page. Use a tool whose redaction feature says that it removes the underlying content, apply the redactions, save under a new name and then check the result. Where no such tool is available, printing the page to an image-only PDF removes the text layer but turns the page into a picture, which has its own costs (no text search, larger file) and should be followed by checks too.

How to check a redacted file

1. Open the file, press select-all and paste into a text editor. Look at what comes out.
2. Search for the words that should be gone.
3. Look at the document properties, bookmarks, comments and attachments.
4. Run an automated check. The PDF Redaction Checker compares what each page shows with the text it contains, reports text hidden under black, white or coloured boxes and under opaque annotations, lists redaction annotations that were not applied, and searches the words you name in page text, annotations, properties, bookmarks and attachment names.

That tool checks for common failure patterns. It does not certify a document, and it cannot see text turned into pictures, leaks through fonts or glyph positions, hidden layers or unusual tricks, so treat a clean result as one input to your own review.

Related tools

To find personal data before you redact, use the PDF Personal Data Scanner; to remove properties, the PDF Metadata Viewer & Remover; to turn forms and annotations into page content, Flatten PDF. For photographs, Face & License Plate Blur changes the pixels themselves.