Can redacted PDFs be made visible again?
The short answer is: it depends on how the redaction was performed. With true redaction — where the text has been physically removed from the PDF — recovery is not possible. With fake redaction — where the text was only visually masked — the original text can be made visible again with simple means. The difference is crucial, particularly in the DSGVO context.
Why some redaction can be reversed
Many people believe that a black bar over text is sufficient to make it illegible. In PDF technology, however, this is not the case. A PDF document consists of several layers: text, graphics and metadata. When you place a black bar over text, the text remains fully intact in the text layer — it is only visually masked.
This means: anyone who opens the PDF in a text editor, removes the graphics layer, or selects and copies the text can read the supposedly redacted text. A sample of publicly indexed PDFs (Essex Software, n=72) found that in roughly every sixth document (approx. 17 %) text remained selectable under the redaction.
True vs. fake redaction
Fake redaction (visual overlay):
- Black bar or rectangle over the text
- Text remains in the PDF content stream
- Copyable via Ctrl+C / Cmd+C
- Visible after removing the graphics layer
- Readable in text editors
True redaction (irreversible removal):
- Text is physically removed from the content stream
- Replaced by an opaque area
- Not copyable, not reconstructable
- Not legible even after removing the graphics layer
- Not present in text editors
The Lower Saxony Data Protection Commissioner explicitly warns: "Often a text passage is only visually masked and can still be read." For DSGVO compliance, only true redaction is sufficient.
How to check whether a redaction is secure
You can test with simple means whether a PDF has been redacted correctly:
Test 1 — Select text: Open the PDF and try to select the redacted area with the mouse. If text is selected, it was only masked, not redacted.
Test 2 — Copy and paste: Copy the contents of the redacted area (Ctrl+C) and paste it into a text editor. If the original text appears, the redaction is insecure.
Test 3 — Open in text editor: Open the PDF in a simple text editor (Notepad, VS Code). Search for known terms that were redacted. If these are found in the text editor, they have not been removed.
Test 4 — Check PDF layers: Professional PDF editors can show and hide individual layers. Hide the graphics layer — if the text underneath becomes visible, the redaction was only visual.
Metadata and hidden risks
Even if the actual text has been correctly redacted, other elements of the PDF can contain personal data:
PDF properties: Fields such as author, title and keywords can contain names or other identifiers.
Comments and annotations: If the document was annotated before redaction, these comments may still be present.
Bookmarks: A well-known case is the AstraZeneca report, where the PDF bookmarks contained the redacted text in plain text — even though the actual redaction had been carried out correctly.
Hidden layers: Some PDFs contain multiple layers that are not immediately visible.
Previous versions: If the PDF was converted from a Word file, metadata of the original file may have been retained.
A complete redaction process must therefore sanitise not only the visible text but also all metadata and hidden information.
How to redact correctly and irreversibly
For DSGVO-compliant redaction, the following steps are recommended:
- Use a professional tool: Use a tool that physically removes the text from the PDF — not just visually masking it.
- Automatic detection: Have personal data detected automatically (regex + NER) instead of searching for it manually.
- Human review: Confirm each proposed redaction individually.
- Sanitise metadata: Ensure that comments, bookmarks and PDF properties are also sanitised.
- Test the result: Carry out the tests described above to verify the irreversible redaction.
- Log: Document the redaction process for audit purposes.
Datenmaske irreversibly removes text from the PDF content stream, sanitises metadata and creates a cryptographically secured redaction log. The NER models are self-hosted on EU servers; no data is sent to US AI services (OpenAI, Google).
FAQ
Can redacted PDFs be made visible again?
With true redaction — where the text has been physically removed from the PDF — recovery is not possible. If the text was only visually masked (e.g. by a black bar), it can be made visible again with simple means.
How do I know whether a redaction is genuine?
Try to select and copy the redacted text (Ctrl+C). Paste it into a text editor — if the original text appears, it was only masked. Alternatively: open the PDF in Notepad and search for known terms.
What is the difference between true and fake redaction?
True redaction physically removes the text from the PDF (irreversible). Fake redaction only places a black bar over the text — the original text remains and can be copied or made visible.
Why were redacted data visible in the AstraZeneca report?
In the AstraZeneca report, the PDF bookmarks were not sanitised. Although the text in the document was correctly redacted, the bookmarks contained the original text in plain text. This shows that metadata is an often-overlooked risk.
Is a black bar in the PDF editor sufficient?
No. A black bar only visually masks the text. The underlying text remains in the PDF and can be made visible with text editors, by copying, or by removing the graphics layer. Such 'redacted' PDFs are not secure.