TOTAL: {[ getCartTotalCost() | currencyFilter ]} Update cart for total shopping_basket Checkout

The latest Council version of the European General Data Protection Regulation (GDPR) provides that personal data may be further processed by the same data controller even if the further purpose is incompatible with the original purpose “if the legitimate interests of that controller or a third party override the interests of the data subject.” The Article 29 Working Party (WP29) and a large number of non-governmental organisations have expressed concerns that this would render the fundamental principle of purpose limitation meaningless and void.

Is this indeed correct­­? We do not think so.

We feel that the approach of the Council is the only feasible way to guarantee protection given that it is much better suited to deal with developments such as the Internet of Things (IoT) and big data.

Let us explain.

The purpose limitation principle consists of two elements:

  • data must be collected for specified, explicit and legitimate purposes only (purpose specification); and
  • data must not be further processed in a way that is incompatible with those purposes (compatible use).

The purpose limitation principle has served as a key principle in data protection for many years. In today’s data-driven society, however, the purpose limitation test has become outdated as a separate test. Imagine a mobile app that on a real-time basis records our health data, predicting how we’ll feel the next day and where to avoid getting the flu. Perhaps pushing the bounds of creepy, this app, however, may be of great value for the World Health Organisation (WHO) to protect civilians from life-threatening infectious diseases.

These two apps collect and use data for the same purpose, namely mapping and predicting health and illness, but our assessment of the two apps is totally different.

The commercial application will not automatically gain societal acceptance, while most of us would see the value of the second application. Whether personal data may be collected and used by such a mobile app is not so much based on the purpose for which the data is collected and processed but on the interest that is served.

As such, we conclude that for data collection and further processing of data, a test based on the legitimate interest is better suited in today’s data-driven society than the current test based on the initial purpose for data collection and whether further processing is compatible with such initial purpose.

Note that with this we do not argue that data limitation or minimisation is no longer relevant.

On the contrary, this principle is still relevant, but in our opinion it should be tied in with the interest served rather than with the original purpose for data collection. This does not mean that more data may be collected and processed. Rather, it may well lead to less data being justifiably processed.

For the record, we are not in favour of a “use-based”" system, in which data collection is left unchecked. Without regulation of data collection, we would end up in a situation where governments and companies would collect rampant data. 

The purpose limitation test has become outdated because, in the past, personal data was primarily a by-product of the purpose for which the data was collected. When we book a flight, we are requested to provide name, address, date of birth and bank account number. The data is a by-product of the service. The requirement of purpose limitation is an objective test to determine what data is justified for that purpose. 

Today, with developments like the IoT, records are no longer a by-product of a purpose, but rather, the data is collected first in order to deliver the service—think smart homes and appliances.

The purpose and the collection of the data then coincide. 

A similar issue arises when a company decides to collect and combine data from public sources in order to analyse the data and provide a new service. It is the company that decides on the service to be provided (on the purpose itself) which then justifies which data is required for such service. As a result, the purpose limitation test is no longer objective and is also non-limiting.

Most online web shops collect online visitor tracking information. While not all data categories collected are strictly necessary to provide access, it is very useful for the website owner to analyse how users behave in his web shop. This data enables the web shop owner to improve its services. Again, the purpose of collection coincides with the interests of the provider to collect the data.

We see these developments implicitly reflected in the opinions of the WP29. Traditionally, the WP29 first applied the purpose limitation test and assessed whether the controller had a legal basis to process the data. In its 2014 Opinion on the Internet of Things, however, the sequence of the tests has been reversed. First it tested whether the controller has a legal basis to process the data (in this case, the legitimate interest ground) and then the purpose limitation test is applied, no more data may be collected and processed than required for the relevant purpose as specified by the controller (which de facto equals the legitimate interest).

If this is how the rules should be applied, then let’s make the system simpler by just applying the legitimate interest test to start with. In this system there will be one test for collection, use, further use and of which test the element of data limitation is already fully part.

Advocates of the purpose limitation principle will point out that this test contains more elements than discussed above, most notably the requirements that the purpose itself must be legitimate and be explicitly specified prior to processing. In its Opinion on the Legitimate Interest Ground, the WP29 indicated that the requirements for a legitimate purpose apply mutatis mutandis to the legitimate interest test, which include the requirement that the purpose must be specified before the collection and processing takes place. Our conclusion is that all elements of the purpose limitation principle are covered by the legitimate interest test. Here the purpose limitation test does not play a role, and therefore only has a role to play in respect of the other legal grounds, such as consent and contract.

Here is the elephant in the room.

Consent and contract are often misused for processing data that would not pass the legitimate interest test. These grounds in themselves do not require the balancing of interests, including the requirement to implement mitigating measures to minimise the impact on the privacy of individuals. In that respect, the legitimate interest ground is better suited and often provides more protection to individuals.

We feel that this is exactly the reason why the WP29 in its Opinion on Purpose Limitation considers relevant for the assessment whether there is a legitimate purpose, also “the general context and facts of the case," including “the nature of the underlying relationship between the controller and the data subjects, whether it be commercial or otherwise.” These factors are not foreseen in the legislative history of the Directive (which refers to an assessment of whether the processing is a violation of the law only). This seems to be an attempt of the WP29 to introduce the contextual analysis of the legitimate interest test also for the other legal grounds. The bottom line is that it results in three tests:

  • whether there is a legitimate purpose;
  • whether there is a legal ground (including the legitimate interest ground), and
  • whether further processing is not incompatible with the initial legitimate purpose.

If the factors listed by the WP29 for each of these three tests are compared, each requires a full assessment of the facts and the context of the case, including the underlying relationship between controller and individual and the reasonable expectations of the individual. Of course, the differences between these tests may be understood by legal experts, but based on our personal experience, both as teachers to practitioners as well as working with companies on their privacy compliance, most people completely miss the point.

The rules are simply too complex and result in an atmosphere of ridiculing the rules instead of an attempt to comply. We are in urgent need of a simpler system. The time has come to recognise the legitimate interest ground as the main legal ground for each and every phase of the data lifecycle. The balancing act of the different interests at stake can then take into the mix whether a contract exists and only allow consent as a mitigating measure rather than predetermined grounds for processing.

Let us remember the important lesson from the First Report on the Directive, which concluded that the Directive was overly strict and complicated and that this is a recipe for non-compliance which in turn undermines the legitimacy of the material norms which these rules aim to protect.

Click here for the extended and annotated version of this article.

photo credit: Qivicon via photopin (license)

1 Comment

If you want to comment on this post, you need to login.

  • comment Simon • Jun 9, 2015
    One pice of law often overlooked in the context of IoT is the Eu e-Privacy directive article 5(3), mistakenly referred to as "cookie-law". Nevertheless, it is broader. Retrieving (any) data from a device connected to a telecom service only is allowed if one of 3 conditions is met; (1) a pure technical reason exists, to facilitate the transmission itself (2) the data is strictly neccessary for a service to the user he requested or (3) unambiguous informed consent.
    As a result his reduces the available legal grounds in GDPR for further process of the data (in case it is personal data) from six to two: "unambiguous consent" or "execution of a contract" (if we interpret "contract" broadly to include the activation of a specific feature or service on the "thing").
    Effectively this means that in IoT space "legitimate interest" can only be a basis for any processing after retrieval if and when the user first has given informed consent to that before any collection started.
    In my view this thwarts the reasoning of the authors of the article regarding the "death of purpose limitation". Unless of course this part of the e-Privacy directive is abolished with the advent of GDPR...