If we learned anything from the surprise election results of 2016, it was that disaffected voters seemed keen to ‘take back control’ from elites. But recent allegations by Carole Cadwalladr for the Observer suggest that voters may have been taking back control to a much lesser extent than they thought. Instead it seems that voters were being subjected to ‘psychological operations’ by data analysis firms such as Cambridge Analytica, linked to pro-Trump and pro-Leave figures. While we were hearing that ‘the public have had enough of experts’ from Michael Gove of Vote Leave, experts in social media manipulation may have been winning his side crucial votes.
Online ‘psychological operations’ are not limited to the specific case of Brexit. Social media marketing has played important roles across various successful elections, including those of Obama in 2008, the Conservatives in 2015, and Trump in 2016. Sophisticated usage of these tools goes beyond using them for broadcasting, and into fine-grained targeting. This works as follows follows: groups such as Cambridge Analytica can use social media data to profile users’ personalities to a worryingly high degree of accuracy; personality profiles, in addition to socio-demographic data, can be used to craft targeted messages which appeal to different types of people; feeding these messages back into social media platforms allows them to reach precisely the sort of ‘persuadable’ people who can be pushed into voting a certain way.
An important question raised by these developments is the extent to which manipulating technology to place personalised, targeted messages into Facebook feeds is the same as manipulating people. Discussions around how ‘free’ our choices really are have been long beloved by both philosophers and advertising executives. But these issues have been further muddied by features inherent in the nature of ‘social’ media. Speaking on Radio 4’s Today programme Giles Kenningham, former head of political press for the Conservatives, appealed to long-standing liberal principles in arguing that ‘people can choose to pick up a paper, and people can choose to engage online’. This ignores an important difference: putting down a newspaper doesn’t disconnect you from your friends, or lead to missed event invites, or generally put you outside of 21st century expectations of socialising. To this we can add economic issues: money has always been able to buy better campaigning, but digital tools can more easily escape traditional limits of material resources, personnel, and national spending limits. Bring all these together, and you end up with some extremely complicated ethical problems.
Engaging with such issues is of great importance to modern democracy. However, such discussions can risk ignoring one crucial psychological phenomenon: people generally don’t like being seen as dupes. Many Vote Leave supporters have, arguably with some justification, resented suggestions that their votes came from ignorance, or from believing bus-borne falsehoods. Further claims that Vote Leave was won by psychological manipulation are unlikely to heal post-Brexit antagonism. That Vote Leave only gained their narrow margin due to democracy being ‘hacked’ may conceivably be numerically correct; but tonally, such claims risk dismissing genuine concerns of voters which must be kept at the heart of political discourse.
The massive political influences of communication technologies, and the hidden networks of people running them, should obviously concern both the public and policymakers. The hairsbreadth closeness of the EU Referendum and Trump results gives credence to the argument that groups with the best social media targeting could have profoundly shaped (and are still shaping) the post-2016 world. And the role of money and personal networks in acquiring better analytics tools should concern anyone who cares for democracy. But this needs to be discussed in a way which doesn’t simply paint a picture of experts – whether those behind our social media feeds, or those trying to expose them – fighting over the minds of malleable voters. Labels such as ‘persuadable’ may be appropriate in psychology laboratories; they’re somewhat uncomfortable when applied to real voters.
We need a language which allows us to acknowledge the powerful psychological role of social media, without repeating divisive messages of awareness versus ignorance. We need to address the risk that international networks of wealthy ideologues could game national electoral systems; but to do so in a way which does not treat voting intentions as straightforward reflections of a Facebook feed. Most importantly we need a digital politics which treats appealing to the undecided voter as a deeply democratic act, listening to and acting upon the interests of those outside the standard political camps. That may be a long way off, or it may not. But in the meantime, digital politics should not become dismissive politics.