futuristic illustration of a young woman width icons für personal data – image AI generated
Information on data protection

The term ‘personal data’ and its evolution

Introduction

The central focus of the General Data Protection Regulation (GDPR) is the processing of personal data. Whilst this may sound self-evident, it repeatedly causes difficulties for many organisations and raises questions. When does data cease to be personal? When does it constitute a category of particularly sensitive data? What is the difference between anonymised and pseudonymised data? Since the GDPR came into force in 2018, the concept of personal data has been shaped by case law, whilst data protection authorities and European data protection institutions have also commented on this concept during this period. Whether data is classified as personal data determines the applicability of the GDPR and entails, amongst other things, information and documentation obligations, as well as the obligation to implement appropriate security measures. Companies should also be familiar with the concept of personal data when assessing the risk to data subjects arising from data processing.

Personal reference

Personal data is defined in Article 4(1), first sentence, of the GDPR as ‘any information relating to an identified or identifiable natural person […]’. According to the established case law of the Court of Justice of the European Union (ECJ), the concept of personal data is to be interpreted broadly. It potentially covers all objective and subjective information relating to the individual in question. This includes, for example, names, contact details, identification numbers, location data and online identifiers. However, whether information is personal in nature always depends on the circumstances of the individual case. 

Case law has further clarified the broadly interpreted concept of personal data in various scenarios. Last year, the ECJ clarified that the context of professional activity in which the data is situated does not preclude it from being personal data. Consequently, the data of the managing director or representative of a legal person is also personal data (ECJ, decision dated 03.04.2025 – Ref. C-710/23). Furthermore, in the view of the ECJ, a statement containing a personal opinion, as an expression of a person’s thoughts, is inevitably closely linked to that person and is therefore always personal data (ECJ, decision dated 04.09.2025 – Ref. C-413/23 P). However, in the view of the Federal Court of Justice (BGH), it is not sufficient for information to be regarded as personal data merely because it relates to a set of facts that has an impact on a specific person (BGH, decision dated 18.12.2025 – Ref. I ZR 115/25). These proceedings concerned the premium history of an insurance contract which, according to the court, was not linked to a specific individual and did not allow any inferences to be drawn regarding the identity or the individual contractual arrangements of the policyholder.

Personal data

Furthermore, it is sufficient if a natural person can be identified indirectly through combination with an identifier. A natural person is regarded as identifiable if they can be identified, directly or indirectly, in particular by reference to an identifier (Article 4(1), second sentence, of the GDPR). As examples of such identifiers, the Regulation cites names, identification numbers, location data, online identifiers or other specific characteristics.

For instance, a member of staff in the HR department can determine which employee corresponds to a specific personnel number. Another example is a vehicle registration number, which can be used to identify the vehicle owner via the vehicle registration authority. If a number plate recognition system has been installed in a car park that uses a camera to capture the number plate of a car entering or leaving the car park, the requirements of the GDPR must be met.

Health data

Article 9(1) of the GDPR lists special categories of personal data which, due to their sensitivity, are subject to special protection. Health data falls within these specially protected categories of data. According to the case law of the ECJ, for data to be classified as health data, it is sufficient if, through logical combination or inference, conclusions can be drawn from the available data regarding the data subject’s past, present or future physical or mental state of health (ECJ, decision dated 14.07.2026 – Ref. C-474/24). In the context of the publication of doping offences committed by sportspeople, such an inference is only possible if the name or category of the prohibited substance or method in question is published.

Last year, the State Data Protection Commissioner explained when the processing of health data is permitted and to what extent employers may gain insight into the state of health of their employees (notice dated 17.07.2025). She states in this regard that employers often labour under a misconception in this respect. The processing of health data to determine an entitlement to continued pay is only permissible insofar as it is actually necessary. The mere suspicion that the condition may be a chronic illness is not sufficient. Rather, there must be a concrete presumption. In principle, consideration should be given to whether less intrusive measures, such as making enquiries with the health insurance fund or consulting a company doctor, could be used.

The processing of health data has implications for the conditions governing the lawfulness of processing and for risk assessment. For example, the Polish Data Protection Authority considers that the mistaken dispatch of a referral confirmation, which contained, amongst other things, information regarding the performance of prenatal diagnostics, posed such a high risk to the patient’s rights and freedoms that the patient should have been notified of the data breach (press release of 27.10.2025).

In addition to health data, Article 9(1) of the GDPR sets out further categories of sensitive data. Financial data are not mentioned there, but by their very nature are also regularly regarded as requiring special protection. 

Technical and organisational measures

There are various technical and organisational measures that remove or restrict the personal reference – including anonymisation and pseudonymisation. These are two distinct processing procedures. Whilst irreversible anonymisation, which makes it impossible to identify the data subject, alters the data in such a way that the personal reference is removed, pseudonymisation reduces the ability to link data to a specific individual without completely removing that link.

Whether and which of these procedures should be carried out in a specific case depends on the risk posed to the data subject by the processing of personal data. The GDPR stipulates that a level of protection appropriate to the respective risk must be ensured for every instance of data processing. The French data protection authority had to assess a case in which a company was conducting studies on diseases and treatment methods using a health database (press release of 28.05.2026). As each patient had been assigned a unique identifier, it was possible to trace their medical history. The authority deemed the risk that patients could be identified to be too high. Consequently, it concluded that not only pseudonymisation but also anonymisation was necessary in this case.

Anonymisation

According to Recital 26 of the GDPR, anonymous data is not subject to the principles of data protection. Anonymisation therefore offers further, low-risk options for using data. Whether data is anonymous must be carefully assessed on a case-by-case basis, and distinguishing it from personal data is often a complex matter. The GDPR does not contain a definition. However, it is generally sufficient that, in practice, re-identification of the data subject cannot be carried out, or can only be carried out with a disproportionate expenditure of time, cost and labour.

The European Data Protection Board (EDPB) published guidelines on anonymisation in July 2026. The guidelines clarify the concept of anonymous data and set out three criteria against which anonymity can be assessed. According to these criteria, data is considered anonymous provided that it is not possible to isolate, link or infer information from the relevant data set.

Furthermore, it must be recognised that anonymisation, as a processing operation in its own right, must be lawful. The anonymity of data is particularly important when external service providers are engaged to process personal data on behalf of and under the instructions of the controller. The data processor may have an interest in further processing the data in anonymised form for its own purposes – for example, as training data for its AI systems or to improve its software. If the anonymisation is carried out by the data processor itself, this in itself constitutes further processing of personal data which does not correspond to the purpose of the data processing on behalf of the controller and requires justification. It is therefore generally advisable to prohibit, as a matter of principle, the further processing of data in anonymised form in the data processing agreement.

Pseudonymisation

Pseudonymisation is described in Article 4(5) of the GDPR as “the processing of personal data in such a manner that the personal data can no longer be attributed to a specific data subject without the use of additional information”. The additional information must be stored separately by the controller.

Last year, the Hanover Regional Court had to address the distinction between anonymisation and pseudonymisation and held that this distinction should be determined from the recipient’s perspective (Hanover Regional Court, decision dated 26.02.2025 – Ref. 128 OWiLG 1/24). Individuals are effectively anonymous insofar as the recipient is unable to attribute the data to a specific person. The court described this process as ‘anonymising pseudonymisation’. If the recipient has no reason to link the pseudonyms to individual persons – for example, if, under the terms of the contract, they could demand the disclosure of the personal reference but are satisfied with the pseudonymised data – the data is, so to speak, anonymous to them.

The ECJ clarified the concept of personal data with regard to the transfer of pseudonymised data to third parties (ECJ, decision dated 04.09.2025 – Ref. C-413/23 P). As a technical measure, pseudonymisation prevents data subjects from being identified by the third party without additional information. Additional information enabling the data to be linked to a specific individual is usually stored by the controller. If the controller transfers the data to a person who has the appropriate means or information to enable identification, non-personal data may become personal data. Therefore, when determining whether data constitutes personal data in the context of pseudonymisation, it depends on whether the third party is actually able to reverse the pseudonymisation and establish a link to a specific individual.

Where personal data is to be sent by email, the encryption requirements vary. In the case of less sensitive data – such as first name, surname and email address – the Düsseldorf Administrative Court considers transport encryption to be sufficient (Düsseldorf Administrative Court, decision dated 02.04.2026 – Ref. 29 K 7351/23). By contrast, according to the Higher Regional Court of Schleswig, end-to-end encryption is the method of choice for sending invoices containing sensitive financial data (Higher Regional Court of Schleswig, decision dated 18.12.2024 – Ref. 12 U 9/24).

Personal data as payment? 

In the case of digital services, consumers sometimes ‘pay’ not with money, but by providing personal data or by consenting to its commercial use. The BGH has referred the question to the ECJ as to whether, in the specific regulatory context, the provision of personal data and consent to its use for commercial purposes are to be regarded as ‘costs’ within the meaning of the European Directive on Unfair Commercial Practices, and whether they must therefore be identified as ‘price’ under the relevant national consumer protection provisions, bearing in mind that personal data may, in principle, constitute contractual consideration, as demonstrated in particular by Section 327(3) of the German Civil Code (BGB), without this prejudging the specific obligation to disclose such information (BGH, decision dated 25.09.2025 – Ref. I ZR 11/20). The ECJ’s ruling is still pending.

The Higher Regional Courts of Cologne and Stuttgart have ruled, with regard to the specific labelling requirement under national consumer law, that the provision of personal data does not constitute a component of the consideration that must be labelled as a ‘price’ (Higher Regional Court of Cologne, decision dated 15.05.2026 – Ref. 6 UKl 3/25; Higher Regional Court of Stuttgart, decision dated 23.09.2025 – Ref. 6 UKl 2/25). In the courts’ view, the term ‘price’ covers only money or a digital representation of value, but not the provision of personal data. However, it does not follow from this that personal data cannot, in general, constitute contractual consideration; this is provided for in principle by section 327(3) of the BGB. The decisions are not yet final in each case and are pending before the BGH.

Conclusion

As soon as personal data is processed, the scope of the GDPR applies. In practice, the concept of personal data raises a number of questions regarding its definition. It is therefore important to monitor current developments, taking into account case law and the activities of data protection authorities.

The term ‘personal data’ must be interpreted broadly and encompasses all objective or subjective information relating to the data subject. It is sufficient for a person to be identifiable by association with a specific identifier. As soon as the available data allows conclusions to be drawn, through mental combination or inference, about the data subject’s past, present or future physical or mental state of health, the data constitutes health data. This category of data falls within the categories of data requiring special protection under Article 9 of the GDPR, which must be taken into account in particular when assessing the lawfulness of data processing, conducting risk assessments and evaluating the adequacy of technical and organisational measures.

Technical and organisational measures also include anonymisation – that is, the removal of any personal reference – and pseudonymisation – that is, the reduction of the ability to link data to a specific individual. According to the criteria of the EDSA, data is only considered anonymous if it is not possible to draw sufficiently reliable conclusions about an individual; to this end, it must not be possible to single out or isolate a person from a data set, nor to link them to other data, and no information about them must be derivable. In the case of pseudonymisation, the decisive factor is whether the respective recipient is actually able to reverse the pseudonymisation and establish a link to a specific individual.