Confidentiality

Anonymisation vs pseudonymisation: the difference under the GDPR, with examples from a legal brief

Replacing your client's name with [PERSONA 1] protects the document, but it does not make it anonymous. Here is the difference, with the text of the GDPR and of the AEPD, an example with a statement of claim and why the facts of the case can still give someone away.

14 November 2026 · 11 min read · By Guillermo Azcona

Pseudonymisation replaces the details that identify a person with tags or codes and keeps the information needed to undo the change separately: the result is still personal data and the GDPR applies. Anonymisation aims to make sure nobody can identify the person by reasonable means, not even whoever did it, and truly anonymous data falls outside the GDPR. A legal brief with [PERSONA 1] and its lookup table is pseudonymised.

What pseudonymisation and anonymisation mean under the GDPR

Pseudonymising separates the data from the person with a key kept elsewhere; anonymising makes sure nobody can join them again. Article 4(5) of the General Data Protection Regulation (GDPR) defines pseudonymisation as:

“the processing of personal data in such a manner that the personal data can no longer be attributed to a specific data subject without the use of additional information, provided that such additional information is kept separately and is subject to technical and organisational measures to ensure that the personal data are not attributed to an identified or identifiable natural person”.

The “additional information” is the key: a table that says who is who, an encryption key or the secret key of a hash. A 2016 guide by the AEPD (Spain's data protection authority, in Spanish) explains it with a minimal example: “if the string AAA replaces a person's name, AAA becomes a key or pseudonym” (our translation). That is why recital 26 of the GDPR says that pseudonymised data “should be considered to be information on an identifiable natural person”.

The GDPR does not define anonymisation in any article. Recital 26 does it, leaving “anonymous information” and “personal data rendered anonymous in such a manner that the data subject is not or no longer identifiable” outside the Regulation. To decide whether someone is identifiable, it says to take into account “all objective factors, such as the costs of and the amount of time required for identification”, with the available technology and its developments.

According to the AEPD, “Anonymisation processing results in a new and unique data set, while pseudonymisation processing results in two new data sets” (“Anonymisation and pseudonymisation”, 6 October 2021): the pseudonymised data and the information that allows it to be undone.

Differences between anonymisation and pseudonymisation

The main difference is whether you can get back to the person: with pseudonymisation you can; with anonymisation, you shouldn't. Everything else follows from that:

PseudonymisationAnonymisation
What is doneIdentifiers are replaced with pseudonyms ([PERSONA 1], a code, a keyed hash) and how to undo it is kept separatelyData is removed or generalised until nobody can identify the person by reasonable means
Can it be undone?Yes, with the additional informationIt shouldn't be, although the AEPD and the European Data Protection Supervisor (EDPS) warn that some anonymisation processes “could be reverted in the future”
Is it personal data?Yes (recital 26), and the GDPR applies to the data, the additional information and the processingNo, if it is truly anonymous (recital 26), although anonymising is itself processing of personal data
Where the GDPR mentions itArticles 4(5), 6(4)(e), 25(1), 32(1)(a), 40(2)(d) and 89(1), and recitals 26, 28, 29 and 78Recital 26
Example in a law firmA brief with [PERSONA 1] and a separate table saying who that isThe number of dismissal cases the firm handled in a year, with nothing about each case

Article 25(1) gives pseudonymisation as an example of data protection by design (“appropriate technical and organisational measures, such as pseudonymisation”) and Article 32(1)(a) as an example of security of processing (“the pseudonymisation and encryption of personal data”). According to recital 28, it “can reduce the risks to the data subjects concerned”, and recital 29 asks for the additional information to be “kept separately”.

A paragraph from a Spanish unfair dismissal claim, with made-up details and translated into English, shows the difference.

Original:

Ms Lucía Garrido Ferrer, holder of DNI 12345678Z and resident at Calle Mayor, 14, Tafalla, has been a shift supervisor at the Pamplona plant of Componentes Arga, S.L. since 2009. She was dismissed on 2 June 2026, after sick leave for a herniated disc and one week after reporting the company to the Labour Inspectorate.

Pseudonymised: direct identifiers become tags, and the same person gets the same tag throughout the document ([PERSONA 1] is Spanish for “person 1”; the DNI is the Spanish national identity number).

Ms [PERSONA 1], holder of DNI [DNI 1] and resident at [DIRECCIÓN 1], Tafalla, has been a shift supervisor at the Pamplona plant of [EMPRESA 1] since 2009. She was dismissed on 2 June 2026, after sick leave for a herniated disc and one week after reporting the company to the Labour Inspectorate.

The lookup table, kept separately, is the “additional information” in Article 4(5):

Real data
[PERSONA 1]Lucía Garrido Ferrer
[DNI 1]12345678Z
[DIRECCIÓN 1]Calle Mayor, 14
[EMPRESA 1]Componentes Arga, S.L.

Anonymous, or so it is meant to be: besides the names, the details that make her recognisable are removed or generalised.

A worker with more than ten years' service at an industrial company was dismissed in 2026, shortly after a period of sick leave and a complaint to the Labour Inspectorate.

With the table, the pseudonymised version turns back into the original; without it, anyone at the plant recognises the shift supervisor from Tafalla dismissed on 2 June. The anonymous version has lost the town, the date, the job, the company and the diagnosis, and it no longer works as a claim. The AEPD and the EDPS warn about this: “The anonymisation techniques will always restrict the ways in which the resulting dataset can be used”.

And even that version is not anonymous for everyone: in the company's HR department there may be only one worker dismissed in 2026 after complaining to the Inspectorate. The 2026 anonymisation guidelines of the European Data Protection Board (EDPB) put it this way: whether data is anonymous “may vary from one entity to another”.

Why a document without names can still identify someone

A document without names still identifies someone when the facts it describes fit only one person. The joint paper by the AEPD and the EDPS “10 misunderstandings related to anonymisation” (April 2021) explains it like this: “While direct identifiers are somewhat trivial to find, indirect identifiers, on the other side, are not always obvious”.

In a legal brief or a case file, these are the indirect identifiers that slip through most often:

Those guidelines read the means of identifying someone broadly, “covering anything from simply reading a document to making a web search”. For datasets, the AEPD proposes k-anonymity (technical note, June 2019), which measures how many people share the same values in the data that could identify them. In a brief, the question is how many people fit what it describes: if the answer is one, the text is not anonymous even without a single name. We cover the techniques for anonymising in anonymising personal data.

A hash or encryption does not anonymise either

Replacing a DNI or a phone number with its hash, or encrypting the document, is pseudonymising: “Encryption is not an anonymisation technique, but it can be a powerful pseudonymisation tool”, says the paper on the ten misunderstandings. The AEPD note «Introducción al hash como técnica de seudonimización de datos personales» (an introduction to hashing as a pseudonymisation technique, October 2019, in Spanish) calculates that with a desktop computer “a dictionary of all the possible hashes of the phone numbers of a given operator can be created in less than 20 seconds” (our translation).

What the EDPB said in 2025 and 2026

The EDPB has put guidelines on pseudonymisation (2025) and on anonymisation (2026) out for consultation:

  1. EDPB Guidelines 01/2025 on pseudonymisation, adopted on 16 January 2025 for public consultation, which closed in March 2025. As of 14 November 2026, the EDPB website only publishes that version 1. According to its press release, pseudonymised data “is therefore still personal data”. The text adds that lookup tables “are personal data” and that, even if the additional information is erased, pseudonymised data “can be considered anonymous only if the conditions for anonymity are met”.
  2. EDPB Guidelines 02/2026 on anonymisation, adopted on 7 July 2026 for public consultation, which closed on 30 October 2026. They propose three criteria to test whether data is anonymous: no record isolation, no linkage and no inference.

What this means for a specific document you send to another firm, an expert witness or an AI depends on the case and on who receives it: ask your data protection officer or a specialist lawyer.

In Spanish: seudonimización or pseudonimización? And is anonimización correct?

In Spanish it is written «seudonimización», with an accent, although «pseudonimización» is also valid. It is the form used in the Spanish text of the GDPR (the English text says “pseudonymisation”) and the one in the RAE's Diccionario panhispánico del español jurídico (its legal dictionary, in Spanish), which defines it with the words of Article 4(5).

The Diccionario de la lengua española lists «seudo-» (“false”) with the variant «pseudo-» (in Spanish), and the Diccionario panhispánico de dudas (in Spanish) notes that in words formed with (p)seudo- “the spellings without p- have already prevailed in many cases (seudología, seudónimo, seudópodo)” (our translation). Seudonimización comes from seudónimo.

«Anonimización» is also correct. The Diccionario de la lengua española includes the verb «anonimizar» (in Spanish), “to express a piece of data relating to entities or persons, removing the reference to their identity” (our translation), although as of 14 November 2026 it has no entry for «anonimización» or «seudonimización». The Diccionario panhispánico del español jurídico does define «anonimización» (in Spanish).

What ablalo's Anonymize section does: pseudonymise

Strictly speaking, ablalo's “Anonymize” section pseudonymises: it replaces data with tags that can be put back later. Its own page says so: “Because the names can be put back with the tag table, strictly speaking it is called pseudonymising”.

With ablalo's Anonymize section you replace names, DNI and NIE, NIG, IBAN, bar number and addresses in a text, a Word file or a PDF, scanned ones included, with tags such as [PERSONA 1], on your computer. “Put the names back” restores the real details in the reply from ChatGPT or another AI. The tag table is not saved: the names are put back before closing. It is built for documents in Spanish, so the tags are in Spanish.

What it doesn't do, and why you should review the result:

For the AEPD and the EDPS, automation helps remove direct identifiers, but “given the importance of the context in the overall process assessment, human expert intervention is needed”. Whether this is enough to comply with the GDPR in a specific processing operation is not something we assess here. Before pasting a brief into an AI, also go through what to ask about where your clients' data is processed.

In short

Pseudonymisation replaces identifiers with pseudonyms and keeps how to undo it separately: the data is still personal data, and the GDPR defines it in Article 4(5). Anonymisation aims to make sure nobody can identify the person by reasonable means, and only then does the data fall outside the GDPR (recital 26). A brief with [PERSONA 1] and its table is pseudonymised; to bring it closer to anonymity you also have to remove or generalise the facts that point to someone, and even then it depends on who reads it.

Sources: Regulation (EU) 2016/679, the General Data Protection Regulation (Articles 4, 6, 9, 25, 32, 40 and 89, and recitals 26, 28, 29 and 78), and its Spanish text in the BOE (in Spanish); AEPD, “Anonymisation and pseudonymisation” (2021), Orientaciones y garantías en los procedimientos de anonimización de datos personales (2016, in Spanish), k-anonymity as a privacy measure (2019) and Introducción al hash como técnica de seudonimización de datos personales (2019, in Spanish); AEPD and EDPS, 10 misunderstandings related to anonymisation (2021); EDPB, Guidelines 01/2025 on Pseudonymisation, version 1, and its press release of 17 January 2025, and Guidelines 02/2026 on Anonymisation, version 1; RAE, Diccionario de la lengua española (seudo- and anonimizar), Diccionario panhispánico de dudas and Diccionario panhispánico del español jurídico (seudonimización and anonimización), all in Spanish. Checked on 14 November 2026.

Frequently asked questions

What is pseudonymisation?

It is processing personal data so that it “can no longer be attributed to a specific data subject without the use of additional information”, kept separately and protected (Article 4(5) of the GDPR). For example, replacing a client's name with [PERSONA 1] and keeping a separate record of who [PERSONA 1] is.

What is the difference between pseudonymisation and anonymisation?

Pseudonymisation can be undone with the additional information, and the data is still personal data. Anonymisation aims to make sure nobody can identify the person by reasonable means, and anonymous data falls outside the GDPR (recital 26).

What is an example of pseudonymisation?

A brief in which the client's name, DNI and address have been replaced with [PERSONA 1], [DNI 1] and [DIRECCIÓN 1], with a separate table saying which person each tag stands for. Replacing a DNI or a phone number with its hash, or encrypting the document, is also pseudonymising.

What does the GDPR say about pseudonymisation?

It defines it in Article 4(5) and gives it as an example of a measure in Article 25(1) (data protection by design) and Article 32(1)(a) (security of processing), among others. Recital 28 says it “can reduce the risks to the data subjects concerned”.

Is it seudonimizacion or seudonimización in Spanish?

With an accent, «seudonimización», because it is a word stressed on the last syllable ending in n. It is the form in the Spanish text of the GDPR and in the RAE's legal dictionary; «pseudonimización» is also valid, because the RAE accepts pseudo- as a variant of seudo-.

Is anonimización correct according to the RAE?

Yes. The Diccionario de la lengua española includes «anonimizar» and the RAE's Diccionario panhispánico del español jurídico defines «anonimización»; as of 14 November 2026, the former still has no entry of its own for the noun.

Tags instead of your client's details, on your computer

ablalo's Anonymize section replaces names, DNI and NIE, NIG, IBAN, bar number and addresses in a text, a Word file or a PDF with tags such as [PERSONA 1], without the document leaving your computer, and puts the names back into the reply from another AI. It does not hide health or trade union membership data: review it before sending it. Built for documents in Spanish. 14-day trial, no card.

See how it works

No card · Windows and Mac · Spanish speech recognition

← Back to the blog