Personal data anonymisation: key to protecting users

Anonymisation is a technique used to dissociate identifying data from the processed records, in order to guarantee the right to data protection

We are living through a period of profound change, in what has come to be known as the data era, in which information is the oil of the 21st century and a key asset in business development. Data underpins business strategies at every level and has great value, which makes it essential to care for it, respect it, and ensure optimal processing that guarantees privacy and data protection for both natural and legal persons.

At this point, anonymisation plays a very important role. As its name suggests, this procedure is the act of eliminating the risk of identifying sensitive data. In other words, these are techniques that enable the responsible use of data, preserving individuals’ anonymity in compliance with current legislation, such as the General Data Protection Regulation (GDPR).

This technique reduces the risks inherent in large-scale data processing by keeping out of public exposure those data that require special protection. However, it does not diminish the ability to use the data and convert it into useful information through technologies such as data science, big data, artificial intelligence, or machine learning, among others.

Anonymisation techniques

Data anonymisation can be carried out using various techniques. Let us look at some of them:

• Homomorphic encryption: allows data to be obfuscated and processed in a concealed manner, as if it were visible. To guarantee the right to individual protection, the protected data can only be made visible with an access key. In other words, only those who have a code to decrypt the data will be able to do so.

• Hash algorithms: this is a mathematical operation that assigns a key to a specific piece of data, without which it cannot be viewed. That key, or fingerprint, represents the real data and without it the data cannot be decoded. The operation is performed in one direction only to provide guarantees to the system: the obfuscated data or microdata always generates the same fingerprint; however, it is not possible to decrypt the anonymised data from that fingerprint.

• Randomisation: as the name suggests, it consists of processing data at random, collecting it for use from the total set of records without identifying who it belongs to, by adding noise (modifying a dataset to make it less accurate), permutation (mixing attributes so they can be attributed to different data subjects), or differential privacy (general processing of data without knowing which person each attribute corresponds to).

• Generalisation: anonymity is guaranteed by modifying the scales and orders of magnitude of the attributes of the natural or legal persons to be protected. To do so, techniques such as Aggregation and k-anonymity (grouping people to prevent their isolation and identification) or l-diversity/t-closeness (similar to the above, adding l differential values to each equivalence class) are used.

What is pseudonymisation?

According to the GDPR, the pseudonymisation of personal data is “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, provided that such additional information is kept separately and is subject to technical and organisational measures to ensure that the personal data are not attributed to an identified or identifiable natural person”.

In other words, it consists of replacing the attributes of the data to be protected with encrypted versions to prevent direct identification, without removing its link to additional information, provided that this is kept separately. An example of pseudonymisation would be replacing customers’ names with alphanumeric codes, so that direct identification is not possible.

Pseudonymisation can be carried out, as in the case of anonymisation, by using hash algorithms, by replacing sensitive data with tokens, or through the use of secret or stored keys.

Differences between anonymisation and pseudonymisation

The basic differences between anonymisation and pseudonymisation are based on two principles: the reversibility of the data and whether it is considered personal data. Let us look at this in more detail:

As we have explained, anonymisation irreversibly alters the data so that the individual to whom it relates cannot be identified, either directly or indirectly. Consequently, under the GDPR, it is no longer considered personal data.

However, pseudonymisation replaces the data to be protected with keys or pseudonyms, which makes it possible to reverse the process. In this case, because it is possible to access the original data again, pseudonymised records continue to be considered personal data.

Examples of anonymisation and pseudonymisation

A clear example of anonymisation can be found in any type of study carried out through surveys. Take, for example, any of the many situation analyses conducted following the outbreak of COVID-19. In these cases, as with the Labour Force Survey, for instance, participants’ data is grouped according to general characteristics such as gender, age ranges, and health status (with or without pre-existing conditions), making it impossible to identify each individual.

On the other hand, if we want an example of pseudonymisation, we can find it in companies’ customer records. To guarantee anonymity, identifying data is replaced with a code that enables internal work and statistical analysis, while ensuring compliance with data protection.