Your cookie preferences

EDPB guidance · Art. 7 GDPR

We use technical cookies essential for the platform to work (login, security, sessions). We would also like to use analytics cookies to understand how to improve it.

You can accept all, reject all, or choose which categories to switch on.Your consent is valid for 6 months and you can withdraw it at any time from the footer.

You will find everything in our privacy notice.

All news
Regulation September 18, 2026 7 min

The AI Act is a retention problem, and almost nobody is treating it as one

Ten years of technical documentation, six months minimum of automatically generated logs, and two clocks running on the same system — while the logs themselves hold personal data about the people the system decided on

TL;DR

Providers of high-risk AI systems keep the technical documentation, quality management documentation, notified body documents and the EU declaration of conformity for ten years from placing on the market or putting into service (Art. 18). Deployers keep the documentation they received for ten years from withdrawal of the system — a different starting point on the same system. Automatically generated logs are kept for a period appropriate to the intended purpose and in any case at least six months (Art. 19 and Art. 26(6)). Serious incidents are reported within fifteen days, ten in the event of death, two for a widespread infringement (Art. 73).

Logging capability is not a retention period

Article 12 requires high-risk systems to technically allow automatic recording of events over the lifetime of the system. It is a design requirement: the capability must exist. How long the records are actually kept is a separate question, answered by Article 19 — a period appropriate to the intended purpose, and not less than six months.

Conflating the two produces the most expensive error available here: keeping every log for the entire life of the system because Article 12 mentions the lifetime. That is disproportionate, it is not required, and on systems that process personal data it is a storage-limitation breach on the largest dataset the organisation holds.

Two clocks on the same system

The provider's ten years run from the moment the system is placed on the market or put into service. The deployer's ten years run from the moment the system is withdrawn. On a system in service for eight years those are two dates almost a decade apart, and an organisation that is both provider and deployer — which is common for internally built systems — runs both.

The provider's obligation is also cumulative rather than rolling: a system updated every quarter for eight years carries its whole documentary history, because the documentation has to show what the system was at each point, not what it is today.

  • Logs of a high-risk system almost always contain personal data about the people the system made decisions on. Two retention regimes then apply to the same file — the AI Act's and the GDPR's — and both purposes belong in the record of processing.
  • Article 73 sets three reporting deadlines, not one: fifteen days as a rule, ten days where a person has died, and two days for a widespread infringement or a serious incident within the meaning of Article 3(49)(b). The clock starts on becoming aware, not on concluding the internal investigation.
  • A model retrained monthly cannot explain a decision taken last year unless the version in use at the time was retained. Keeping the model — or at least its parameters and description — therefore becomes a retention period in its own right, and it is one nobody writes down.
  • The documentation is kept at the disposal of national competent authorities. That is availability on request, not archival storage: a backup nobody can read within a reasonable time does not discharge the obligation.

One row for “AI system” is not a record

The technical documentation, the automatically generated logs, the post-market monitoring data and the incident reports have four different purposes, four different periods and, for the logs, two legal regimes at once. A record of processing with a single line for the system describes none of them. The row that most often goes missing is the model version itself, because nobody thinks of a set of weights as a document — until somebody asks why a decision was taken.

The practical test is the same one that works for accounting records: does the retention column hold a number, or does it say which artefact follows which provision? If it holds a number, it is describing a simplification. Here the simplification is newer than the others, which means nobody has yet been caught by it — not that it is safe.

Looking for a workspace for your DPO work?

DPO Workspace is built by a certified DPO. 30-day free trial.

Start free

Related articles

Regulation
nullathe decision on a worker taken by the algorithm alone, from 23 October

A dismissal decided by the algorithm is void: from 23 October in Italy

Published in Official Gazette no. 234 of 8 October 2026 and in force since 23 October, decree no. 179/2026 provides that final decisions on the employment relationship cannot be based solely on automated processing, that the worker may request an understandable explanation, and that decisions taken in breach are void. In parallel, AI systems affecting work organisation enter the risk assessment under Article 28 of decree 81/2008. For a DPO this is the rare case where the consequence is not economic but civil, and it changes how the check is framed.

Oct 08, 2026 8 min
Regulation
3signals that make notification due

Poland's UODO: not every breach must be notified. Public-register data on one side, PESEL, health and financial data on the other

If an invoicing system leaks only a sole trader's name, business name, tax number and registered address, already public in the business register, risk is normally unlikely. If it also leaks the national ID number, purchase details revealing health, or financial data, notification is due. Either way, the decision goes into the internal register.

Oct 05, 2026 5 min
Regulation
120days to charge you, not to fine you

The 120 days are not the deadline to fine you: they are the deadline to charge you

Anyone who built an appeal on the authority's delay should re-read the calendar. The Court clarifies three things: the complaint procedure and the sanctioning procedure are autonomous, so overrunning the nine-to-twelve months on the complaint does not extinguish the power to fine; the 120 days are for notifying the charge, not for imposing the fine; and they run from the «definitive finding», which is not the first notice but the moment the authority can assess every element — including the hearing of your side. After the charge, only the five-year limitation period remains.

Sep 20, 2026 7 min