As the US election approaches its terrifying final act, our social media, TV screens and office chit chat seem to be entirely monopolised by the Battle of Clinton and Trump. The crescendo of scandal, drama and improbability has swept everyone along – over the past few months I’ve caught up with industry folk from every continent and sooner or later the conversation has turned to the US election. So perhaps that’s why I can’t help but connect absolutely everything else that happens with next week’s world-changing vote.
In particular, this week’s story about UK insurance brand and Facebook had my mind going to sinister places.
If you haven’t seen the story, Admiral had planned to use Facebook data to analyse young drivers’ posts for indications of recklessness or a sensible disposition and adjust quotes accordingly. The first thing to note is that the information they were looking for and the correlations with driving ability are flaky at best. Overuse of exclamation marks, for example, was singled out as a signal of overconfidence and not, for example, a half-assed attempt to convey confidence or enthusiasm on social media when your heart really isn’t in it. So for one thing the pseudoscience and overgeneralisations are just intellectually offensive.
But more sinister, the app also shows how easily impersonal algorithms can comb our online data and the potential for abuse. Imagine a future where your financial wellbeing is linked to your syntax or punctuation preferences? That’s just a few clicks away from Black Mirror’s Nosedive episode, right? Fortunately Facebook put their foot down just before the project was due to go live – Facebook’s privacy rules mean that user data cannot be used to determine eligibility for offers or products.
It did get me thinking though. Over the past few months, Facebook has been clogged with people voicing their political opinions, sharing political articles and propagating pieces of political content. What if brands could use that data to determine eligibility or to decide who to offer discounts to?
Sure, brand planners might include general voting preferences as they develop a picture of a target market – but as soon as businesses start drilling into specific peoples’ specific political standpoints and using that to make decisions (say, whether or not to grant a mortgage), things become fundamentally undemocratic. It’s manipulative – share the ‘correct’ opinion or risk losing out financially or being excluded entirely.
That’s why voter privacy and freedom of association are so important.
You could argue that slathering your political beliefs all over social media for all to see means that you forfeit the right to protect that data from businesses… but I’d argue that these platforms are how we share ideas and campaign and communicate now and that genie is not going back into the bottle.
When it comes to targeting, through programmatic advertising or eligibility algorithms, I would love to see social platforms and brands treat individuals’ political data as sacred, untouchable. It may sound unworkable, but then if democracy is to survive the coming information age it must be protected.