Anonymize student data for AI
The need to anonymize student data for AI arises in primary and secondary schools and universities: a teacher who wants to write report card comments, plan remedial work from essays or analyze a class's results handles files that name minors or students, with their INE (the French national student identifier), their school platform (ENT) login or their Parcoursup number (the French higher education admissions platform). This data falls under the GDPR and, for minors, calls for extra care. Biffer processes these files on the teacher's computer, with no connection, and replaces each identifier with a stable alias.
This page describes the files concerned, the applicable rules, the AI uses and the fields of the teacher profile.
Lists, report cards, essays, files: a teacher's files
The files of a teacher or a school are named by design: they are used to follow pupils one by one.
- Class lists and exports (Excel, CSV): last name, first name, date of birth, INE, pupil number, ENT login, contact details of parents or guardians.
- Report cards and comments (Word, PDF): grades, averages, comments per subject, enrollment number.
- Essays and digital assignments (Word, text, PDF): name in the header, sometimes details of personal life in a piece of writing.
- Guidance and Parcoursup files: Parcoursup number, personal statement, Fiche Avenir (the school's assessment form), class council opinion.
- Meeting reports: class councils, educational team meetings, individual support plans, safeguarding reports.
Biffer processes PDF, Word, Excel, CSV and text files while preserving the formatting: a grade sheet stays a table, a report card keeps its structure. Essays scanned as images go through built-in optical character recognition, offline, then are redacted by blacking out the pixels, with no restoration; for handwriting, the preview lets you check and complete the recognition. The general method is described on the page prepare a document for AI.
Student data and the GDPR: the school's obligations
A school or university is the controller for its pupils' and students' data under the GDPR and the French Data Protection Act. Preparing a pupil file before submitting it to an online AI service protects the school, which keeps control over what it entrusts to a service outside its contracts. Recital 38 of the GDPR also states that children merit specific protection with regard to their personal data.
On top of this come professional secrecy for some information (health, family situation, safeguarding reports) and the duty of discretion of French public servants. Three points deserve attention.
- The INE, the pupil number or the ENT login identify a pupil as surely as their name, and follow the pupil throughout their schooling.
- A pupil's comments and writing are data about them as soon as the file makes it possible to find them.
- Where processing happens matters: with offline software, files stay inside the school and the data flow remains entirely under your control.
Biffer prepares, you validate: the teacher reviews the preview and the school has the final word on what it shares.
AI use cases and the Biffer workflow
Once pseudonymized, a teacher's files lend themselves well to AI assistance, because teaching is about the work rather than the pupil's identity.
- Write report card comments from brief notes on each pupil.
- Analyze a CSV export of grades to spot pupils struggling with a skill.
- Suggest remedial work from a batch of digital essays.
- Summarize the report of a class council or an educational team meeting.
The workflow is the same every time.
- You drop the file into Biffer with the teacher profile enabled.
- The local analysis highlights people, dates of birth, addresses, emails, phone numbers, INE, pupil, student, Parcoursup and enrollment numbers, and ENT logins.
- You untick false positives (the name of the school, an author being studied) and add what is missing, all occurrences at once.
- The document comes out with stable aliases (ELEVE_001 for a pupil, INE_001, PERSONNE_004; the alias labels are in French); in Word and PDF, redacted data is highlighted in black, an option enabled by default.
- The mapping table, encrypted or in plain text, restores the real names in the comments written by the AI, ready to be copied over.
The queue processes a batch of essays or report cards with the same aliases from one to the next. Biffer is free for five files per month, after a one-time email confirmation in the software; beyond that, Biffer for Mac or Biffer for PC costs €10 excl. VAT (€12 incl. VAT) per computer, paid once: unlimited files, lifetime license tied to the computer, all redaction features included (see the pricing page). The steps are detailed on the features page.
The fields the teacher profile recognizes
The teacher profile adds school and higher education identifiers to the general detections. Here is the exact list, as it appears in the software.
| Field | Where you find it |
|---|---|
| INE number | Pupil exports, certificates of enrollment, school records |
| Pupil number | Class lists, school management software, report cards |
| Student number | Transcripts, student cards, attendance sheets |
| Parcoursup number | Application files, Fiches Avenir, opinions |
| Enrollment number | Enrollment files, exams, competitive exams |
| ENT login | Platform exports, accounts, login logs |
On top of these come the four fields shared by all professions (file number, reference, contract number, employee ID) and the general detections: people, emails, phone numbers, postal addresses, French company numbers (SIREN and SIRET), French social security number, IBAN, license plates, dates. For an identifier specific to your school, the custom keyword field described on the professions page does the rest. The trainer profile can be added if needed. Biffer is published by Romain Rissoan, trainer, consultant and founder of OPCADIA, in Lyon.
Frequently asked questions
Are first names on their own, common in essays and comments, detected?
Biffer detects people in several forms: first and last name, title followed by a name, and first name alone when it appears in its dictionary of common first names. In a class list, the first name and last name pair is recognized reliably. In an essay where the pupil writes only their first name, or in a comment about "Lucas" with no surname, detection works for usual first names; for a rare first name, the preview lets you add it once, for all its occurrences in the file and in the queue. The dictionary reflects first names common in France.
Can a teacher use Biffer on their personal computer?
Yes, it is even the most common case. Biffer installs on Mac or PC and works with no connection: files stay on the teacher's computer, from import to export, in their hands alone. Biffer is free for five files per month, all redaction features included, after a one-time email confirmation in the software, with no password. For higher volumes, Biffer for Mac or Biffer for PC costs €10 excl. VAT (€12 incl. VAT), paid by card (Stripe): a lifetime license for the version purchased, tied to that computer, unlimited files. On a personal computer, you apply your school's rules on processing pupil data, and Biffer helps you do so.
How do you copy the comments written by the AI into the right report card?
That is the purpose of the mapping table. At export, Biffer offers to save it, encrypted (Argon2id and AES-256-GCM) or in plain text, depending on how well protected your computer is. You give the AI a table where each pupil is ELEVE_001, ELEVE_002, with their grades and your brief remarks. The AI returns comments written for each alias. You load this text back into Biffer with the table: the real names are restored, locally, and you copy each comment into the school management software after review. Without a table, the pseudonymization is irreversible, which suits a statistical analysis of the class.
Your business documents anonymized, ready to be published or handed to an AI tool.
Biffer runs on your computer without sending a single file. Free for five files a month, unlimited for €10 excl. VAT per computer.
100% offline · Free: 5 files a month · No file sent