How to Demonstrate the Irreversibility of Data Anonymisation? – Another Expert Lecture Completed

On 12 June 2026, an open expert lecture was held as part of the series "Data Protection and Technology", organised by the Personal Data Protection Office in cooperation with the Social Team of Experts at the President of the Personal Data Protection Office.

The event was opened by Mirosław Wróblewski, President of the Personal Data Protection Office, and the lecture entitled "A One-Way Ticket: How to Demonstrate the Irreversibility of Anonymisation in AI and Data Science Projects?" was delivered by Mariola Więckowska, MSc Eng, a member of the Social Team of Experts at the President of the Personal Data Protection Office.

During the meeting, participants learned about four key areas related to the anonymisation and pseudonymisation of personal data in projects involving artificial intelligence:

· pitfalls of pseudonymisation in AI models,

· synthetic data as a privacy-enhancing tool,

· accountability workshop before the President of the Personal Data Protection Office,

· PETs metrics, re-identification tests and DPO checklists.

"Confusing data security with legal anonymisation is the most common mistake made by IT architects," emphasised Mariola Więckowska.

Referring to EDPB Opinion 28/2024 on certain aspects of data protection related to the processing of personal data in the context of AI models, the expert explained that if the probability of recovering data from model parameters is higher than negligible, the entire model should fall under the GDPR.

During the lecture, Mariola Więckowska discussed, among other things, the business and legal benefits of using synthetic data, its advantages and disadvantages, and the role of the Data Protection Officer in implementing the principle of privacy by design.

With regard to the anonymisation process, the speaker stressed that it cannot be approached in a static manner: "It is not the case that we anonymise and forget. Tests must be carried out regularly to ensure that the standard is maintained."

At the end, she presented a checklist that can be used by a DPO. It was prepared on the basis of EDPB guidelines and related materials provided by ENISA and the Personal Data Protection Office.

After the lecture, a Q&A session took place, during which participants could obtain additional clarifications on the topics discussed.

The lecture was addressed, among others, to Data Protection Officers, data controllers, information security specialists, and individuals responsible for developing and implementing solutions based on artificial intelligence.

Those interested in personal data protection are encouraged to take part in future open expert lectures organised as part of the "Data Protection and Technology" series.