Spanish Supervisory Authority and EDPS release guidance on hashing for data pseudonymization and anonymization purposes

On November 4, 2019, the Spanish Supervisory Authority (“AEPD”), in collaboration with the European Data Protection Supervisor, published guidance on the use of hashing techniques for pseudonymization and anonymization purposes. In particular, the guidance analyses what factors increase the probability of re-identifying hashed messages.

The AEPD explains that the probability of re-identification increases if more information is available on the hash values used (e.g., that they were created on the basis of Spanish phone numbers of a certain operator). The guidance provides examples of how controllers can make the re-identification of hashed messages more difficult. These examples include encrypting the message (prior to hashing), encrypting the hash value, or adding “salt” or “noise” (i.e., a random number) to the original message.

According to the AEPD, the use of hashing techniques for pseudonymization and anonymization purposes requires companies to analyze the risk of re-identification, taking into account the hashing technique used. The risk analysis must assess the hashing process and all the other related elements, such as the information that the controller retains about the hash value after the hashing (e.g., that the hash values consist of Spanish phone numbers). The analysis should lead to an objective evaluation of the probability of re-identification of the hashed message over time.

The guidance also lists a number of “basic” considerations when using hashing for anonymization or pseudonymization purposes, such as ensuring secure access to the hashing process and periodically auditing the management processes of the hashing system.

Finally, according to the guidance, in order for a hashing technique to be considered an anonymization technique, the risk analysis must—in addition to the above considerations—assess two factors:

• whether information which permits the re-identification of the hashed message has been deleted; and
• whether the applied hashing technique will remain sufficiently robust over time.

Note that earlier this year the AEPD also released guidance on applying K-anonymization to data sets.

Spanish Supervisory Authority and EDPS release guidance on hashing for data pseudonymization and anonymization purposes posted first on http://ronenkurzfeld.blogspot.com