As Theresa May prepares to trigger Article 50 and take the UK out of the EU, there has been a lot of talk about “the clear will of the people” as expressed through the referendum last year. But as reports emerge about the use of big data and social media by both Leave.EU and Vote Leave, we really need to ask whether elections run with this kind of tactic can really be said to reflect the free will of the electorate. With these techniques being linked to the Trump campaign win in the United States and the possibility of their use in upcoming elections across Europe, the future of liberal democracy as we know it may well depend on how we answer that question.
Freedom of thought is a fundamental right protected under international law. Unlike many other rights that allow for restriction in certain circumstances, there is an absolute right to think what you like in the “forum internum” or the inner space of your mind. This inviolable freedom has been described as “the foundation of democratic society” and “the basis and origin of all other rights”. Crucially, freedom of thought includes the freedom to keep our thoughts private – we may not be coerced into revealing our thoughts. And it is connected to the corresponding right to freedom of expression and opinion which provides the social backdrop crucial to critical and intellectual thought. But one of the most important aspects of freedom of thought is freedom from coercion in our thoughts. This includes freedom from indoctrination or influence on our conscious or subconscious mind through manipulation. Freedom of thought has received little attention in the courts, perhaps because of an assumption that our inner thoughts are beyond the effective scope of state intervention. But as technology and science develop ever clearer pictures of the way our thought processes work and how they can be altered, it is time for this assumption to be challenged.
In 2012, Facebook published the results of research it had conducted on 700,000 Facebook users showing how it could alter the emotional state of users by manipulating their news feeds so that they received an abnormally low number of either negative or positive posts. Researchers found those shown more negative comments posted more negative comments and vice versa. There was a backlash against the research with users complaining about manipulation, but although Facebook apologised, it simply said the research should have been “done differently”. In 2015, a Cambridge University study showed that psychological profiling based on Facebook likes allowed researchers more insight into a user’s personality than their close friends and family. In 2016, a big data company claimed to have brought these two concepts together in a technique called “Behavioural Microtargeting” which was used for the Trump campaign in the US and, according to some reports, for the leave.EU campaign in the EU referendum in the UK.
Although the effectiveness of the techniques has been questioned and one of the companies highlighted says their techniques don’t give them “special powers over people”, their description of their services does raise questions about freedom of thought. Regardless of whether or not they were able to influence the outcome of elections so far – are we really happy with the development of this type of behavioural micro-targeting and its potential impact on our thoughts, behaviour and democracy? There are arguments that all campaigning and advertising is designed to alter our behaviour and the way we think. But this type of technique is different because it aims to access our thoughts on an individual level without our knowledge and use that information to change our thoughts, emotional states, opinions and therefore our voting behaviour through “dark posts” that we are unaware of as political advertising. This goes to the heart of freedom of thought.
In the UK, the Information Commissioner’s Office has started to investigate the use of big data in the Brexit referendum from a data protection perspective. Differences in data protection laws in the US and the EU may explain why this is happening here but not in the US. But the bigger question is whether this type of behavioural micro-targeting, particularly in a political context, should be allowed at all because of its goal to interfere with freedom of thought on a fundamental level. The real meaning of interference with freedom of thought is starting to become tangible – if we still believe it is an absolute right and the foundation of democracy, it is time to take urgent action to protect it.
If our freedom of thought is compromised, calls to respect the “will of the people” sound hollow. Governments are under an obligation to protect our human rights from interference by third parties. Regardless of what may have happened in the US or in the EU referendum, there is an urgent need to take action and set up a clear legal framework to protect our freedom of thought and the integrity of our democracies in the digital age.
 (Nolan v Russia, App. 2512/04 para 61)
 Rene Cassin, France, as reported in Martin Scheinin, Art. 18, in: A. Eide et al. (eds.) UDHR: A Commentary. Scandinavian University Press 1992, 266