Last updated: 21 November 2026
Quick answer: In the pseudonymisation vs anonymisation debate, the key difference is reversibility. Pseudonymised data can still be linked back to a person using separately held additional information, so it remains personal data for the controller and the GDPR applies. Properly anonymised data can no longer identify anyone by means reasonably likely to be used, so the GDPR no longer applies.
DPOs hear the two terms used interchangeably by product teams, researchers and vendors. The legal consequences could hardly be more different: one keeps data inside the GDPR, the other takes it outside.
This guide explains both concepts, the EU and UK tests, the impact of the 2025 CJEU ruling in the SRB case, and a quick checklist for deciding which label applies.
What is the difference in pseudonymisation vs anonymisation?
Pseudonymisation replaces or removes direct identifiers but keeps a way to re-identify; anonymisation removes the possibility of identification altogether.
The GDPR defines pseudonymisation in Article 4(5) as processing personal data so that it can no longer be attributed to a specific data subject without additional information, provided that information is kept separately and protected by technical and organisational measures. Recital 26 confirms that pseudonymised data which could be attributed to a person using additional information should be considered information on an identifiable person.
Anonymisation is not defined in an article, but Recital 26 states that data protection principles do not apply to anonymous information. To decide whether someone is identifiable, you must consider all the means reasonably likely to be used, taking account of objective factors such as cost, time and available technology.
| Aspect | Pseudonymisation | Anonymisation |
|---|---|---|
| Reversible? | Yes, with the additional information | No, by any means reasonably likely to be used |
| GDPR status | Personal data for the controller | Outside the GDPR once achieved |
| Typical techniques | Tokenisation, keyed hashing, coded IDs | Aggregation, generalisation, noise, suppression |
| Legal role | A security and data minimisation safeguard (Articles 25, 32, 89) | An outcome that ends GDPR obligations for that data |
| Main risk | Key or lookup table compromise | Re-identification through linkage or inference |
Why does pseudonymisation still matter if data remains personal?
Because the GDPR rewards it. Pseudonymisation is named as an example of data protection by design in Article 25 and as a security measure in Article 32(1)(a). It supports safeguards for research and statistics under Article 89 and can be relevant when assessing breach risk.
The EDPB’s Guidelines 01/2025 on pseudonymisation, published for consultation in January 2025, explain how to apply it in practice. A central idea is the “pseudonymisation domain”: the set of people or systems from whom the link to individuals is meant to be hidden. The safeguard only works if the additional information stays out of that domain.
In practice: record pseudonymisation as a control in your DPIAs and your record of processing activities, and name who holds the keys. That shows accountability and helps when a regulator asks how risk was reduced.
What did the CJEU decide in EDPS v SRB?
On 4 September 2025, the Court of Justice ruled in Case C-413/23 P, EDPS v SRB that pseudonymised data is not necessarily personal data for every person in every case. The case concerned Regulation 2018/1725, which mirrors the GDPR for EU institutions.
Three points matter for DPOs:
- Recipient perspective. If technical and organisational measures genuinely prevent a recipient from attributing the data to individuals, the data may not be personal data for that recipient.
- Controller perspective. For the original controller, which holds the additional information, the data remains personal data.
- Transparency is judged at collection. The Court held that identifiability for the duty to inform about recipients must be assessed from the controller’s point of view at the time of collection, so the controller still had to tell data subjects about the recipient.
Common mistake: reading the ruling as “pseudonymised data is anonymous”. It is not. The ruling makes status relative to who holds what, and it does not relieve the disclosing controller of its obligations.
How do the EU and UK approaches compare?
The core tests are closely aligned, but guidance and reform timelines differ.
| Topic | EU | UK |
|---|---|---|
| Legal test | GDPR Article 4(5) and Recital 26, means reasonably likely to be used | UK GDPR equivalent provisions |
| Main guidance | EDPB Guidelines 01/2025 on pseudonymisation | ICO anonymisation guidance, including the motivated intruder test |
| Case law | CJEU EDPS v SRB, 4 September 2025 | Not binding in the UK, but may be persuasive |
| Reform | Digital Omnibus proposal on relative identifiability, under negotiation | Guidance under review after the Data (Use and Access) Act 2025 |
The ICO’s anonymisation guidance describes identifiability as a spectrum and uses the “motivated intruder” test: could a reasonably competent person with no prior knowledge, but with access to public resources, re-identify someone? The ICO notes the guidance is under review because of the Data (Use and Access) Act 2025.
In the EU, the European Commission’s Digital Omnibus, proposed on 19 November 2025, would state that pseudonymised data is not personal data for an entity without the means to re-identify. The European Parliament’s legislative train shows it is still going through the legislative process.
If you want to document these decisions consistently across processing activities, start your 14-day free trial.
How do you decide whether data is anonymised?
Run this checklist and record the answers.
- Have all direct identifiers been removed, and is any lookup table or key destroyed or truly out of reach?
- Could records be singled out, linked with other data sets or used to infer information about a person? These are the three risks the Article 29 Working Party identified in its 2014 opinion on anonymisation techniques.
- Who will receive the data, and what other information could they reasonably access?
- Would a motivated intruder succeed with public data and available technology?
- Have you scheduled a periodic re-assessment as technology and available data change?
If you cannot answer confidently, treat the data as pseudonymised personal data and apply GDPR controls.
Key takeaways
- Pseudonymised data remains personal data for the controller that can re-identify it.
- Anonymised data falls outside the GDPR, but the bar is high and must be reassessed over time.
- The CJEU’s SRB ruling makes identifiability relative to the recipient, without removing the discloser’s duties.
- Record pseudonymisation as a safeguard in DPIAs, ROPAs and processor contracts.
Frequently asked questions
Is hashed data anonymous under the GDPR?
Usually not. Hashing an identifier such as an email address creates a consistent value that can often be matched against a list of known inputs, especially without a secret key. In most cases, hashed data is pseudonymised personal data. Only when the result cannot reasonably be linked to a person by anyone likely to access it could it be considered anonymous.
Do we still need a lawful basis to anonymise data?
Yes. Anonymisation is itself processing of personal data, so it needs a lawful basis and must respect the other GDPR principles. Once data is effectively anonymised, the output falls outside the GDPR. Document the anonymisation method and your reasoning, so you can show why the output is no longer personal data.
Does pseudonymisation reduce breach notification duties?
It can. If pseudonymised data is breached but the additional information remains secure, the risk to individuals may be lower, which affects whether you must notify the supervisory authority or individuals. It does not remove the need to assess and record the breach, and you must still consider the specific circumstances.
Can we share pseudonymised data with a vendor without GDPR obligations?
Not automatically. After the SRB ruling, the data may not be personal data for a vendor that genuinely cannot re-identify it. But you, as the disclosing controller, still hold the key, so your own obligations continue, including transparency about recipients. Assess the vendor’s position case by case and document the safeguards.
Conclusion: label data by what can be done with it
The pseudonymisation vs anonymisation question comes down to who can re-identify people and by what means. Treat anything reversible as personal data, reserve “anonymous” for outputs you have tested, and document the decision. For related work, read our DPIA guide, and use DPIA, ROPA and vendor risk tools in Enactia’s AI-powered GRC platform. Our GDPR framework page covers the wider regulation.
To discuss your approach, contact us or book a demo or start your 14-day free trial.
This article is for general information and is not legal advice.
