Anonymization or Pseudonymization? The GDPR Difference That Matters

The difference is simpler than it sounds. Anonymization is final: it removes the link to a person for good, and the link cannot be rebuilt. Under Recital 26 of the GDPR, such data no longer falls under the law. Pseudonymization only replaces data with reversible placeholders. With the right key, the originals can be restored. That is why pseudonymized data stays personal data, and the GDPR still applies. ONYRI Sanitize pseudonymizes: it replaces sensitive values with reversible tokens. That lowers your risk, but it does not remove your legal duties.

The short difference in one sentence

Both methods protect data, but only one takes you out of the GDPR. Keep these three points in mind:

• Anonymization: permanent, not reversible. The link to the person is gone. The GDPR no longer applies.
• Pseudonymization: reversible with a key. The link to the person remains. The GDPR still applies.
• Rule of thumb: if you can still trace the person in theory, it is pseudonymization, not anonymization.

Anonymization: the link to the person disappears for good

Anonymization means no one can link the data back to a person, not even you. There is no key, no lookup table, no path back. True anonymization is strict: even combining the data with other sources must not make the person visible again. In practice, this is hard to achieve.

• Example: aggregated statistics such as '4,200 customers in Hamburg use plan A'. No single person is visible.
• Example: a survey stripped of all names, IDs, and rare traits, so that no path back is possible.
• Result: under Recital 26, anonymous data falls outside the GDPR. It can be analyzed freely.
• Caution: too few traits removed (age, city, and job together) can make the person identifiable again. Then it is not anonymous.

Pseudonymization: protected, but still personal data

Pseudonymization replaces identifiable details with placeholders. For example, 'Anna Weber' becomes '[NAME_1]'. The link to the person is not deleted, only set aside. Whoever holds the key can turn '[NAME_1]' back into 'Anna Weber'. That is exactly why this data stays personal data. The GDPR names pseudonymization as a safeguard, not as an exit from the law.

• Definition: Art. 4(5) GDPR describes pseudonymization as processing where data can no longer be linked without extra information.
• Safeguard: Art. 32 GDPR lists pseudonymization as an example of an appropriate technical security measure.
• Legal status: under Recital 26, pseudonymized data stays personal data. The GDPR applies in full.
• Benefit: risk drops clearly. Anyone who sees the pseudonymized data without the key cannot identify the person.

Anonymization vs. pseudonymization, side by side

This table sums up the three questions that matter day to day: Is it reversible? Does it stay personal data? Does the GDPR apply?

What ONYRI Sanitize does, and what it does not

ONYRI pseudonymizes; it does not anonymize. Before your text goes to an AI such as ChatGPT, Copilot, Gemini, Claude, or Le Chat, the tokenizer detects sensitive details and replaces them with reversible tokens. The architecture matters: the tokenizer runs fully in your browser. The mapping from token to original value never leaves the browser. When you use the built-in Chat, only the already-pseudonymized (tokenized) text is sent to the model. The answer is restored locally.

• Honest framing: pseudonymization reduces data exposure and supports data minimisation under Art. 5 GDPR.
• Equally honest: pseudonymized data stays personal data. ONYRI does not guarantee GDPR compliance and does not make your output anonymous.
• In practice: fewer real names, account numbers, or API keys in the prompt mean less risk if something goes wrong at the provider.
• Limit: your legal duties (a lawful basis, processing agreements, informing data subjects) still stand.

When does each method fit?

The choice depends on your goal. If you want to remove data from data protection law for good, you need true anonymization. If you want to work with data daily and still keep the link, pseudonymization is the right path.

1. 1Goal, lasting analysis without a personal link: anonymization, provided no path back is truly possible.
2. 2Goal, daily work with AI tools: pseudonymization, so you can restore the real values later.
3. 3Goal, lower risk without losing the case: pseudonymization as a safeguard under Art. 32 GDPR.
4. 4In every case: check whether the measure is enough in your context, and keep meeting your GDPR duties.