Cambridge Analytica, a company that uses “data modeling and psychographic profiling” (according to its website), is credited with having decisively contributed to the outcome of the presidential election in the U.S.. They did so by using “a hyper-targeted psychological approach” allowing them to see trends among voters that no one else saw and thus to model the speech of the candidate to resonate with those trends. According to Mashable, the same company also assisted the Leave. EU campaign that leaded to Brexit.
How do they do it?
“We collect up to 5,000 data points on over 220 million Americans, and use more than 100 data variables to model target audience groups and predict the behavior of like-minded people” (my emphasis), states their website (for comparison, the US has a 324 million population). They further explain that “when you go beneath the surface and learn what people really care about you can create fully integrated engagement strategies that connect with every person at the individual level” (my emphasis).
According to Mashable, the company “uses a psychological approach to polling, harvesting billions of data from social media, credit card histories, voting records, consumer data, purchase history, supermarket loyalty schemes, phone calls, field operatives, Facebook surveys and TV watching habits“. This data “is bought or licensed from brokers or sourced from social media”.
(For a person who dedicated their professional life to personal data protection this sounds chilling.)
Under US privacy law this kind of practice seems to have no legal implications, as it doesn’t involve processing by any authority of the state, it’s not a matter of consumer protection and it doesn’t seem to fall, prima facie, under any piece of the piecemeal legislation dealing with personal data in the U.S. (please correct me if I’m wrong).
Under EU data protection law, this practice would raise a series of serious questions (see below), without even getting into the debate of whether this sort of intimate profiling would also breach the right to private life as protected by Article 7 of the EU Charter of Fundamental Rights and Article 8 of the European Convention of Human Rights (the right to personal data protection and the right to private life are protected separately in the EU legal order). Put it simple, the right to data protection enshrines the “rules of the road” (safeguards) for data that is being processed on a lawful ground, while the right to private life protects the inner private sphere of a person altogether, meaning that it can prohibit the unjustified interferences in the person’s private life. This post will only look at mass psychological profiling from the data protection perspective.
Does EU data protection law apply to the political profilers targeting US voters?
But why would EU data protection law even be applicable to a company creating profiles of 220 million Americans? Surprisingly, EU data protection law could indeed be relevant in this case, if it turns out that the company carrying out the profiling is based in the UK (London-based), as several websites claim in their articles (here, here and here).
Under Article 4(1)(a) of Directive 95/46, the national provisions adopted pursuant to the directive shall apply “where the processing is carried out in the context of the activities of an establishment of the controller on the territory of the Member State“. Therefore, the territorial application of Directive 95/46 is triggered by the place of establishment of the controller. Moreover, Recital 18 of the Directive’s Preamble explains that “in order to ensure that individuals are not deprived of the protection to which they are entitled under this Directive, any processing of personal data in the Community (EU – n.) must be carried out in accordance with the law of one of the Member States” and that “in this connection, processing carried out under the responsibility of a controller who is established in a Member State should be governed by the law of that State” (see also CJEU Case C-230/14 Weltimmo, paras. 24, 25, 26).
There are, therefore, no exceptions to applying EU data protection rules to any processing of personal data that is carried out under the responsibility of a controller established in a Member State. Is it relevant here whether the data subjects are not European citizens, and whether they would not even be physically located within Europe? The answer is probably in the negative. Directive 95/46 provides that the data subjects it protects are “identified or identifiable natural persons“, without differentiating them based on their nationality. Neither does the Directive link its application to any territorial factor concerning the data subjects. Moreover, according to Article 8 of the EU Charter of Fundamental Rights, “everyone has the right to the protection of personal data concerning him or her”.
I must emphasise here that the Court of Justice of the EU is the only authority that can interpret EU law in a binding manner and that until the Court decides how to interpret EU law in a specific case, we can only engage in argumentative exercises. If the interpretation proposed above would be found to have some merit, it would indeed be somewhat ironic to have the data of 220 million Americans protected by EU data protection rules.
What safeguards do persons have against psychological profiling for political purposes?
This kind of psychological profiling for political purposes would raise a number of serious questions. First of all, there is the question of whether this processing operation involves processing of “special categories of data”. According to Article 8(1) of Directive 95/46, “Member States shall prohibit the processing of personal data revealing racial or ethnic origin, political opinions, religious or philosophical beliefs, trade-union membership, and the processing of data concerning health or sex life.” There are several exceptions to this prohibition, of which only two would conceivably be applicable to this kind of profiling:
- if the data subject has given his explicit consent to the processing of those data (letter a) or
- the processing relates to data which are manifestly made public by the data subject (letter e).
In order for this kind of psychological profiling to be lawful, the controller must obtain explicit consent to process all the points of data used for every person profiled. Or the controller must only use those data points that were manifestly made public by a person.
Moreover, under Article 15(1) of Directive 95/46, the person has the right “not to be subject to a decision which produces legal effects concerning him or significantly affects him and which is based solely on automated processing of data intended to evaluate certain personal aspects relating to him, such as his performance at work, creditworthiness, reliability, conduct, etc.”. It is of course to be interpreted to what extent psychological profiling for political purposes produces legal effects or significantly affects the person.
Another problem concerns the obligation of the controller to inform every person concerned that this kind of profiling is taking place (Articles 10 and 11 of Directive 95/46) and to give them details about the identity of the controller, the purposes of the processing and all the personal data that is being processed. In addition, the person should be informed that he or she has the right to ask for a copy of the data the controller holds about him or her and the right to ask for the erasure of that data if it was processed unlawfully (Article 12 of Directive 95/46).
Significantly, the person has the right to opt-out of a processing operation, at any time, without giving reasons, if that data is being processed for the purposes of direct marketing (Article 14(b) of Directive 95/46). For instance, in the UK, the supervisory authority – the Information Commissioner’s Office, issued Guidance for political campaigns in 2014 and gave the example of “a telephone call which seeks an individual’s opinions in order to use that data to identify those people likely to support the political party or referendum campaign at a future date in order to target them with marketing” as constituting direct marketing.
- The analysis of how EU data protection law is relevant for this kind of profiling would be more poignant if it would be made under the General Data Protection Regulation, which will become applicable on 25 May 2018 and which has a special provision for profiling.
- The biggest ever fine issued by the supervisory authority in the UK is 350.000 pounds, this year. Under the GDPR, breaches of data protection rules will lead to fines up to 20 million euro or 4% of the controller’s global annual turnover for the previous year, whichever is higher.
- If any company based in the UK used this kind of psychological profiling and micro-targeting for the Brexit campaign, that processing operation would undoubtedly fall under the rules of EU data protection law. This stands true of any analytics company that provides these services to political parties anywhere in the EU using personal data of EU persons. Perhaps this is a good time to revisit the discussion we had at CPDP2016 on political behavioural targeting (who would have thought the topic will gain so much momentum this year?)
- I wonder if data protection rules should be the only “wall (?)” between this sort of targeted-political-message-generating campaign profiling and the outcome of democratic elections.
- Talking about ethics, data protection and big data together is becoming more urgent everyday.
Find what you’re reading useful? Consider supporting pdpecho.