Redacting documents for Woo requests

Verilak proposes which personal data to redact, you decide on every proposal, and the text under the black box is then really gone. With a legal ground on every redaction and an inventory list for the decision under the Dutch Open Government Act (Woo).

Why redaction often goes wrong

A black bar in a PDF editor or in Word covers the text, but does not always remove it. Whoever opens the document can often simply select and copy the text. The search index, fonts or an earlier version inside the file can also contain the redacted words. That is how names and citizen service numbers have ended up in published Woo documents more than once.

Verilak solves this by turning a redacted page into an image. Nothing is left under the box. A separate check then searches the whole file for any remaining trace of the redacted text.

From case file to publication in four steps

  1. Make proposals

    Verilak searches all documents in the case file for personal data and makes a proposal for every finding.

  2. Review

    A staff member looks at every proposal on the page, chooses to redact or not and picks the legal ground. Nothing is redacted without review.

  3. Apply

    Verilak redacts the approved boxes with real removal and puts the ground on the box.

  4. Check and publish

    An independent redaction check, a before and after comparison and a Woo inventory list per case file.

What Verilak detects

Detection uses check digits where possible, so you get fewer false proposals.

Numbers

  • Citizen service number (BSN), with the eleven test
  • IBAN, with check digits
  • Credit card numbers, with Luhn
  • Case numbers and licence plates

Contact details

  • Email addresses
  • Phone numbers
  • Postcodes
  • IP addresses

Names and sensitive data

  • Names, via salutation, initials, greeting and signature
  • Special categories of personal data (GDPR articles 9 and 10)
  • Your own word lists
  • Search term across the whole case file

Images and scans

  • Faces in images
  • Scanned pages via OCR
  • Also pages that were saved rotated
  • Draw redactions yourself, with mouse or keyboard

Built for the Open Government Act

  • Legal ground on every redaction. Choose the ground from the Woo, for example respect for privacy. The ground is shown on the box.
  • Woo inventory list. A list of all documents per case file, with what was redacted and on what ground.
  • Consistent across the case file. A name that is redacted in one document is also proposed in the other documents of the same case file.
  • Merge with numbering. Merge the documents into one PDF with page numbers and document numbers.
  • Searchable again. After redaction, OCR makes the remaining text selectable and searchable again.
  • Accountability. Every decision goes into the audit log with a hash chain, exportable as PDF/A.

Verilak never decides by itself. Detection makes proposals; a staff member reviews them and is responsible for the decision. That keeps the human review the Woo and the GDPR require.

Try redaction with your own case files?

In a six week pilot you test Verilak with real Woo case files, on your own workstations.