How to Anonymize Resumes for Blind Hiring
To anonymize a resume for blind hiring, remove everything that identifies the candidate or hints at who they are, namely name, photo, address, email, phone number, date of birth, nationality and links to personal profiles, and keep what shows whether they can do the job: skills, experience, responsibilities and results. The quickest reliable way is to replace each identifier with a neutral tag such as [Person 1] rather than a black box, so the resume stays readable, and to keep a private table that maps each tag back to the candidate for the interview stage.
What to remove and what to keep
| On the resume | Remove? | Why |
|---|---|---|
| Name, including in the header, file name and email address | Yes | Signals gender and ethnic background |
| Photo | Yes | Signals age, gender, ethnicity, appearance |
| Home address and postcode | Yes | Signals neighbourhood and background; city alone is usually fine |
| Email, phone, LinkedIn and portfolio links | Yes | Lead straight to the name and photo |
| Date of birth, age, marital status, nationality | Yes | Protected characteristics |
| Graduation years | Consider | A proxy for age |
| School and university names | Your policy | Some employers hide them to reduce prestige bias |
| Employer names | Usually keep | Relevant to experience; hide if a referee or small firm identifies the person |
| Hobbies and memberships | Consider | Can reveal religion, politics, family status |
| Skills, job titles, responsibilities, achievements | No | This is what screening should be based on |
| Gaps in employment | No, but brief reviewers | Often reflect caring responsibilities |
Black boxes or tags?
A resume covered in black bars is hard to read and still leaks information: a reviewer can see how long a name is and where it appears. Pseudonymization replaces each identifier with a consistent tag instead. The candidate becomes [Person 1] everywhere in the document, a referee becomes [Person 2], the home address becomes [Address 1]. The text still reads as sentences, and nothing identifying is left underneath because the original words are removed from the file. The difference between the two approaches is explained in pseudonymization vs anonymization.
Step by step
- Decide your rules once, in writing: which fields come off every resume, and at what stage names are revealed. Apply the same rules to every candidate for the role.
- Give each application a candidate number and use it as the file name. The original file name usually contains the candidate's name.
- Upload the resumes to Redact PDF AI's anonymize mode. PDFs, scans and photos are accepted; convert Word files to PDF first.
- Choose what becomes a tag. People, addresses, emails and phone numbers should. If you want reviewers to see employers and dates of employment, leave organizations and dates unticked; if your policy hides them, tick them.
- Check each resume in the editor. Add a box over the photo, which is an image rather than text and so is not detected automatically, and over anything the AI could not know is identifying, such as a personal website written out in the text.
- Download the anonymized PDF and the correspondence table. The table lists each tag with the original value.
- Give reviewers only the anonymized PDFs. Store the tables where the screening panel cannot open them.
- Reveal names after shortlisting, by matching the candidate number to the original application.
Limits worth knowing
- Tags restart in each document. Every resume has its own
[Person 1]. The candidate number in the file name, not the tag, is what tells applications apart. - Context can still identify. "Captain of the national under-21 hockey team" narrows the field to one person. Blind screening reduces bias; it does not make a candidate impossible to recognise.
- Hiding dates has a cost. With dates as tags, reviewers can no longer see how long each job lasted. Many teams keep employment dates and remove only the date of birth and graduation years by hand.
- It only covers the screening stage. Interviews are not blind. The point is that the decision about whom to interview is made on skills and experience alone.
- This is pseudonymization, not anonymization. Whoever holds the correspondence table can re-identify every candidate, so the resumes remain personal data under GDPR. Processing takes place in the EU and Switzerland, and documents are never used to train AI models.
Using an AI assistant to screen or summarise?
If you plan to paste applications into ChatGPT, Claude or a similar assistant, anonymize them first: once the original is uploaded, the provider has every candidate's details. The method is in how to anonymize a document before using ChatGPT.
Frequently asked questions
What is blind hiring? Blind hiring is screening job applications with identifying details removed, so that the decision to interview is based on skills and experience rather than on a name, photo, age or address.
What information should be removed from a resume for blind recruitment? Name, photo, home address, email, phone number, links to personal profiles, date of birth, nationality and marital status. Many employers also remove graduation years, and some remove school names.
Is an anonymized resume still personal data? Yes, as long as someone can link it back to the candidate, which your HR team can. Treat anonymized resumes with the same care, retention periods and access rules as the originals.
How do we find out who the candidate is after screening? Through the candidate number on the file, which matches the original application, or through the correspondence table downloaded with each anonymized resume.
Does it remove the candidate's photo? Not automatically. A photo is an image, so add a box over it in the editor; it takes a few seconds per resume.
Can it handle resumes in other languages? Yes. Text is read in more than 100 languages and names are detected in about 80. Tags are written in English, French, German, Spanish, Italian, Dutch or Portuguese, matching the document's language where it is one of those, and in English otherwise.
Should employer names be hidden too? It depends on your policy. Employer names help reviewers judge experience, so most teams keep them. Hide them when the employer is so small or specific that it identifies the person.
Get started
Anonymize a resume with tags and see the result before you roll it out to a whole vacancy. Responding to a staff member's data request instead? See how to redact a subject access request.