<\/p>\n
Anna Flanagan<\/a> and Malcolm Dowden<\/a>\u00a0of Pinsent Masons, who specialise in data protection law, were commenting after the European Data Protection Board (EDPB) adopted new guidelines on anonymisation<\/a> (34-page \/ 1.37MB PDF).<\/p>\n The processing of personal data is subject to data protection law. In the EU, the General Data Protection Regulation (GDPR) imposes a range of conditions on personal data processing, with the potential for significant financial penalties to be levied on organisations that do not comply with those conditions.<\/p>\n In some cases, organisations can achieve their objectives using data that has been stripped of identifiers and no longer qualifies as personal data. They might seek to remove information from which individuals can be identified from data sets, to reduce restrictions and compliance burdens affecting how they can use and share that data. This can also reduce risks to individuals.<\/p>\n Anonymised data falls outside the protection of the GDPR. However, data that is thought to have been fully anonymised might only have been “pseudonymised” or masked, meaning that it is still in scope of the law. The EDPB\u2019s new guidance sets out its view of the practical tests to be applied to determine whether data is genuinely anonymised.<\/p>\n In GDPR terms, data will be anonymous only where it no longer relates to an identified or identifiable natural person, taking account of the means reasonably likely to be used by the relevant party to identify individuals. The EU\u2019s highest court ruled last year that data can be considered anonymised in the hands of one party but not necessarily in the hands of another \u2013 it depends whether the recipient of data has the means to identify someone from the data<\/a>. The EDPB said it is therefore \u201cnecessary to consider the different perspectives of the relevant entities\u201d when performing the test for anonymity.<\/p>\n \u201cWrongly believing that personal data has been anonymised, only to find that re-identification is reasonably possible, is a common cause of GDPR infringement,\u201d said Flanagan. \u201cIf data is wrongly treated as anonymous, the GDPR compliance measures that should accompany personal data processing, such as transparency information, lawful-basis analysis, data protection impact assessment where required, and appropriate security controls, may be absent or inadequate.\u201d<\/p>\n Dowden added: \u201cThe EDPB’s new guidance provides a structured and practical test. Following its steps and documenting reasoned conclusions will materially reduce the risk of breach.\u201d<\/p>\n The EDPB has set out a three-step test to help businesses understand whether information is truly anonymised or whether there is still a risk of re-identification.<\/p>\n According to the EDPB, data should be treated as anonymous only where there is no realistic risk of record isolation, linkage or inference. In practice, this means assessing whether the data: contains a unique combination of attribute values that relate to a single individual; contains an individual\u2019s record that could be linked, with certainty or high likelihood, to another record relating to the same individual in a different dataset; or allows specific and meaningful inferences to be drawn about an individual. The EDPB frames this as the \u2018no record isolation, no linkage, no inference\u2019 test, and its guidance provides detailed explanations and examples to help organisations apply those criteria in practice.<\/p>\n \u201cThe framework can be applied in two different ways: one which considers the differences in capabilities between those who might identify the data subject (\u2018the contextual approach\u2019) and one which does not (\u2018the simplified approach\u2019),\u201d the EDPB said.<\/p>\n \u201cThe contextual approach reflects the full nuances of the legal standard for anonymisation and allows the controller to assess whether data is anonymous for each relevant entity based on their respective capabilities. The simplified approach, on the other hand, can go beyond the legal standard and may lead an anonymising controller to treat data as though it is not anonymous even if it would actually be so for some relevant entities. However, it also offers a more convenient option for controllers who choose to use it, provides greater confidence that data is actually anonymous and can be combined with the contextualised approach to refine the findings,\u201d it said.<\/p>\n