One of the most insightful, and scary, stories around data use in the past 6 months was ‘The great British Brexit robbery: how our democracy was hijacked’ by Carole Cadwalladr for the Guardian/Observer.
The premise of the article, which may be slightly biased but still relatively accurate, is that through the use of personal data collected across a vast amount of areas, ‘The company [Cambridge Analytica] also (perfectly legally) bought consumer datasets – on everything from magazine subscriptions to airline travel’ combined with ‘psych data’, a few people managed to influence the election of Donald Trump and the Brexit referendum result.
According to a source who previously worked at Cambridge Analytica, one of the main protagonists in the article, named as ‘Paul’ they then used this data and combined it with military knowledge to persuade people to act in a certain way - ‘That’s what it is. Psyops. Psychological operations – the same methods the military use to effect mass sentiment change. It’s what they mean by winning ‘hearts and minds’.’
It is a really interesting take on how data is now being used to influence world events by just a few very well resourced people. One of the key themes that allows this to happen is that the rules around the use of certain types of data are so easy to bypass, mainly because they don’t really exist.
It is an issue for every major democracy, with election rules being written slowly over decades to cover new media as it progressed at a comparatively slow pace. It meant that new media like TV, text message, and to some extent online banner ads, could easily fall within these remits with slight tweaks to existing rules. After all, a party may run 10 variations of a banner ad which could be quickly checked by various official bodies to ascertain fairness and accuracy. However, with the speed that data has spread, enabling with it hyper-individualized targeting, these same rules are redundant.
Where rules were originally based on the idea that a few messages were being sent to millions of people, we now have a situation where thousands of messages are being sent to millions of people designed to elicit particular reactions. According to Cadwalladr’s Guardian article, in the case of Cambridge Analytica, the company at the centre of the Brexit and Trump election controversies, these messages were deliberately using military psychological techniques combined with this targeting. In the article Cadwalladr noted that:
‘Finding “persuadable” voters is key for any campaign and with its treasure trove of data, Cambridge Analytica could target people high in neuroticism, for example, with images of immigrants “swamping” the country. The key is finding emotional triggers for each individual voter.’
With both Brexit and Trump, the margins were incredibly fine between victory and defeat, with Trump even losing the popular vote by over 3 million and less than 1% of the electorate being the difference in Brexit. It means that these kinds of actions are having a profound impact on the eventual results. Given the partisan nature of a huge range of issues today, not only in politics, but across the board, this grey area of influence have become essential to victory.
So with these actions going largely unchecked, is it possible for governments or rule setting bodies to create regulations suitable for the modern data-driven messaging environment?
It is clearly a big challenge given the almost unfathomable amount of messaging that is sent out in this way, but if these enforcement bodies can utilize similar technologies to help identify this messaging then it’s a good start. This may not be so far off either, with many of the social platforms beginning to build political teams within their organizations, which will help to create messaging for political campaigns and organizations. For instance, Facebook has recently begun to recruit veteran political operatives from major political parties with the Guardian reporting some of these as, ‘Rishi Saha, a former head of digital communications at Downing Street; Karim Palant, a former chief policy adviser to Balls, Labour’s shadow chancellor, and Theo Lomas, a former political consultant for Crosby Textor, the PR firm of the man running the Tories’ 2017 general election campaign.’
With a more formal politically focused team that works directly with parties and organizations representing them, a trail will be created with more easily trackable data usage, messaging, and spend. If Facebook is going to create a political team, it would also not be much of a jump to think they will also be tracking political messages sent without the use of this team. This creates the opportunity for legal bodies to set broader rules for parties about the messaging and data being utilized, with more specific rules and data use audits for the platforms through which these messages are sent.
With the overseer bodies having the ability to audit how voters are targeted and the messaging used to target them, this data can be used to reverse engineer profiles that fit these rules, allowing them to create monitoring profiles to check messaging as it is sent out.
However, this type of political messaging is moving swiftly. After the 2012 campaign where the Obama team were rightly praised for their use of data, it has taken a more complex turn, turning a potentially positive development into a grey area where the rules are unclear. Once these rules surrounding data use on social platforms are put in place there will no doubt by another development that will have a profound impact. When that comes it will need to be dealt with, but the one thing that needs to be done today is to have rules set around the current systematic use of data, something that is clearly easier said that done.