Hidden metadata in Word and PDF files: what it reveals and how to remove it
A document is more than what you see on screen. Every Word file, PDF or photo carries information nobody typed on purpose: who created it, which software produced it, when it was edited, which comments were made during review, what text was deleted with track changes, and where a picture was taken. This is hidden metadata, and in documents that leave your organisation it can reveal far more than you meant to share.
The UK regulator takes this seriously. In July 2025 the ICO published dedicated guidance on disclosing documents to the public securely, prompted by serious breaches in which hidden data in published spreadsheets exposed personal information. It covers hidden rows, columns and worksheets, active filters and metadata, and applies whether you are publishing online, answering a subject access request or a Freedom of Information request. For a formal definition, see our glossary entry on metadata.
What Word, PDF and image metadata can reveal
| File | What it can contain | Risk |
|---|---|---|
| Word (.docx) | Author, last modified by, company, template, editing time, revision number | Identifies staff and suppliers; exposes another client's template |
| Word (.docx) | Comments, tracked changes, hidden text, versions, custom XML | Internal opinions, figures and names you thought were deleted |
| Word (.docx) | Headers, footers, watermarks, embedded objects (a whole Excel sheet) | Data you don't see when reading the body text |
| Document information (title, author, subject, keywords, creator application, producer, dates) | Drafter's name, software and real creation date | |
| XMP metadata, attachments, annotations, hidden layers, form fields, incremental saves | Earlier versions and invisible content | |
| Photos (JPG, HEIC) | EXIF: date, camera model and serial number, GPS coordinates, thumbnail | Location of homes or individuals |
One case surprises many people: when you convert a Word document to PDF, the Word title and author are copied into the PDF properties. And a photo with GPS data pasted into a Word file may, depending on how it was inserted, keep its EXIF data inside the final PDF.
How to check a file's metadata
- Word: File > Info shows properties (author, last modified, title) and warns if comments or tracked changes remain. Check for Issues > Inspect Document lists everything the Document Inspector finds.
- PDF: in Adobe Acrobat or Reader, File > Properties (Ctrl+D) opens the Description tab with title, author, subject, keywords, application and producer; the additional metadata button shows XMP. Check the comments and attachments panels too.
- Images: in Windows, right-click > Properties > Details shows camera, date and, if present, latitude and longitude. On a phone, the photo's info panel usually shows a map.
- In bulk: command-line utilities such as ExifTool are widely used by technical teams to review many files at once.
How to remove it
Word: Inspect Document
Microsoft's Document Inspector finds comments, revisions, versions, ink annotations, document properties and personal information, custom XML, headers, footers, watermarks, hidden text and invisible content. Save a copy first (removed data may not come back, even with Undo), run File > Info > Check for Issues > Inspect Document, and choose Remove All for the relevant categories. Two limits matter: it only detects text formatted as hidden, not white text on a white background, and it does not find objects covered by other objects. The ICO guidance recommends tools like this but also warns against relying on ineffective techniques such as changing text colour.
Before inspecting, accept or reject all tracked changes and delete comments deliberately, so you know what they said. When saving as PDF, open Options and untick "Document properties" if you don't want the title and author carried into the PDF.
PDF: properties and sanitising
You can clear the fields in File > Properties by hand, but that leaves XMP, attachments, layers and incremental versions untouched. Acrobat Pro includes Sanitize Document (and selective removal of hidden information), which strips metadata, comments, hidden layers and other content and saves a new file. Without it, printing to PDF produces a file without layers or attachments but usually adds fresh properties and loses selectable text, accessibility tags and signatures.
Photos
In Windows, Properties > Details > Remove Properties and Personal Information creates a clean copy. On phones, turn off camera location if you don't need it and use the share-without-location option where available.
Why it matters under UK GDPR
Data protection law doesn't mention metadata by name, but the data minimisation principle (Article 5(1)(c)) and the security principle (Article 5(1)(f)) both apply. Publishing an author's name or a photo's location when it isn't needed, or releasing a spreadsheet whose hidden sheet contains personal data, can be a personal data breach. The ICO's guidance suggests practical safeguards: choose a simple disclosure format where possible (for example exporting data to CSV, which only carries visible values), use inspection tools, and keep a record of the redactions and transformations you apply.
Before you send or publish a document
- Accept or reject tracked changes and delete comments.
- Run the Document Inspector in Word, Excel or PowerPoint.
- Export to PDF without document properties if they aren't needed.
- Review and clear the final PDF's properties and XMP.
- Remove EXIF and GPS data from included photos.
- Check the result by opening the final file's properties.
Cleaning metadata with Anonimatum
When you have many files or mixed formats, the manual process becomes fragile. Anonimatum includes a tool to remove metadata from PDF, Office and image files: it handles PDF, DOCX, XLSX, PPTX, JPG, PNG and WebP, takes a single file or a whole folder, and returns the clean file or a ZIP. When you anonymise a Word document, the output also comes without comments, with tracked changes accepted and with metadata stripped, so the clean-up happens in the same step.
Metadata is only one way data escapes. If you're also redacting visible content, read about the PDF redaction mistakes that leak data: a black box drawn over text does not delete it.
Frequently asked questions
Does saving as PDF remove Word metadata?
Not entirely. The title and author are copied into the PDF properties unless you untick "Document properties" in the save options, so check the result.
Does the Document Inspector catch everything?
No. It only finds text formatted as hidden, not text coloured to match the background, and it misses objects covered by other objects. A visual review is still needed.
Do social networks and messaging apps strip photo GPS data?
Some platforms remove it on upload and others don't, and behaviour can change. The safest approach is to remove location data yourself before sharing.