Skip to content
‹ The story: ST-008

Matter file · ST-008

EDPB guidelines open for consultation

What do the draft guidelines on anonymisation, web scraping and GDPR fines say, and when do comments close?

As at 26 October 2026

Introduction#

The European Data Protection Board has three sets of draft guidelines out for comment. Comments on anonymisation and on web scraping close at 23:59 CET on 30 October 2026. Comments on when to impose a fine close at 23:59 CET on 13 November 2026. All three are drafts, not final guidance.

How it got here#

7 July 2026. The EDPB adopted Guidelines 02/2026 on anonymisation and Guidelines 03/2026 on web scraping for generative AI, both for public consultation.

17 September 2026. It adopted Guidelines 04/2026 on when a supervisory authority should impose a fine, for public consultation.

What each draft says#

Anonymisation (02/2026). Data is anonymous when it does not relate to an identified or identifiable person. The EDPB says each relevant entity should judge anonymity from its own point of view, taking account of the means reasonably likely to be used. A firm can apply the test in a “contextual” way or a “simplified” way. It has three criteria: no record isolation, no linkage and no inference. If data meets all three, the EDPB states a firm can safely treat it as anonymous. If it fails any, the firm needs more analysis.

Who it touches: firms that anonymise customer, transaction or case data for analytics, sharing or retention.

Web scraping (03/2026). The draft covers private entities that scrape personal data from the internet, themselves or through a contractor, to train generative AI. The GDPR applies where scraping processes personal data. Controller and processor roles must be analysed case by case. To keep collection to a minimum, the EDPB suggests:

  • considering synthetic data;
  • setting precise collection criteria;
  • filtering out some categories of data;
  • excluding websites that hold such data or clearly object to scraping.

It also sets out the three conditions, all of which must be met, for relying on legitimate interest under Article 6(1)(f) GDPR.

Who it touches: firms that use scraped open-source data to build screening, adverse-media or fraud models.

When to fine (04/2026). The draft sets a five-step method for deciding whether to fine, as well as or instead of other corrective measures under Article 58(2) GDPR:

  1. Can the infringement lead to a fine?
  2. Can this party be fined for it?
  3. Was it committed intentionally or negligently?
  4. Do the Article 83(2) factors make it minor? If so, as a rule, no fine. If not, the EDPB states there is a strong presumption of a fine.
  5. Would a fine be effective, proportionate and dissuasive under Article 83(1)?

The draft adds to the EDPB's earlier guidelines on calculating fines.

What comes next#

Short term (to January 2027)

  • 30 October 2026, 23:59 CET. Comments close on anonymisation and web scraping.
  • 13 November 2026, 23:59 CET. Comments close on when to fine.

No date yet

  • Final versions of all three guidelines.

Open questions#

  1. Will the final anonymisation guidelines keep the three criteria and the two ways of applying them?
  2. Will the final fines guidelines keep the strong presumption of a fine once an infringement is not minor?

What this is not#

  • None of the three is final guidance. Each is a consultation draft.
  • The fines guidelines do not set how much to fine. They decide whether to fine, alongside the earlier guidelines on calculation.

Reading#

Primary

Official

Changes to this file

  1. 26 October 2026 · First published.

Updated in place. Where the matter stood on each earlier date is in the story’s position ledger. Open the story ›