There is something creepy about being stalked on “social media”; creepier still that we’re unaware of the stalking; creepiest of all that our minds – and our votes – are being insidiously influenced by the stalker.
This is what lies at the heart of the furore over the data firm Cambridge Analytica and the allegation that it obtained personal data from Facebook users to target political messages.
There should be no question that potential breaches of data privacy law should be investigated and, were those breaches to be proven, the culprits punished.
Millions of innocent Facebook users would think long and hard about using this platform if they felt their personal details and preferences were being obtained by third parties and the data extracted to target them to effect an election vote or political outcome.
Regulatory authorities in the US, the UK and continental Europe are now mounting investigations that may lead to greater scrutiny of Facebook. That many may have already acted to close their accounts has added to the damage the affair has caused – wiping some $50 billion off the stock market value of shares in the company.
But it is one thing to ensure regulated access to personal data on the world wide web. It is something else to infer that the alleged breaches of access influenced the election of Donald Trump in the US and the ‘Leave’ vote in the UK referendum on membership of the EU. That is the widespread assertion. But how do we know such targeting had any tangible influence?
Cambridge Analytics has not helped its cause by arguing simultaneously that claims of such an effect on the outcome are absurd while boasting of its influence at the highest level of the Trump campaign.
More generally it has brought critical scrutiny to bear on political activity on social media outlets generally and how effective they are in influencing our behaviour. Millions of Amazon customers will be well used to targeted advertising and promotions based on our previous searches and purchases reminders of items to which we may only have given a scant perusal.
That may seem innocent enough – until we wonder how much of our previous personal internet shopping history is stored in a digital library that can harbour more information about our personal predilections than we have chosen to reveal to relatives and friends. Altogether more concerning is that our personal political preferences and beliefs can be inferred from other seemingly unrelated personal data. If we are known to purchase certain styles of clothing or cosmetic brands, can it be deduced that we are somehow more likely to vote Left or Right? And what of the messages that are then sent to our inboxes? How can they be shown to have had an influence on the way we vote? Little has been revealed so far as to the content of the messages sent out to targeted US voters.
Here, too, our interpretation of information can reflect our personal bias. For some, this controversy over undue influence is straightforward. Voters in the US presidential election did not choose rationally; rather they were manipulated by external forces who made use of social and/or psychological assessments to target them in a certain way.
The explanation finds understandable favour among Democrat supporters of Hillary Clinton as to why she lost against Trump. Was not the presidency stolen from her? Did not Cambridge Analytica track down ‘gullible’ voters and play on their gullibility? But similar stratagems were used in the Obama campaigns.
It cannot be said to be a sympathetic interpretation of democracy that voters can be played in this way. But then ‘hidden persuasion’ has been a fact of life since modern advertising began. Pre-dating this controversy for more than 50 years was Vance Packard’s best-selling book ‘The Hidden Persuaders’. “Large-scale efforts are being made, often with impressive success”, he wrote, “to channel our unthinking habits, our purchasing decisions, and our thought processes by the use of insights gleaned from psychiatry and the social sciences. Typically these efforts take place beneath our level of awareness, so that the appeals which move us are often, in a sense, ‘hidden’.”
Debate has raged ever since as to how effective – if at all – such efforts have been as influential as Packard argued. For were he right, there is hardly an advertising campaign that could have failed. But many do.
Among the more nuanced versions of this approach is that target messaging is not efficacious in persuasion – hidden or otherwise – than in reinforcing existing beliefs.
John Thornhill, writing in the Financial Times earlier this week, quoted a study published in the journal Science based on thousands of ‘information cascades’ on Twitter between 2006 and 2017. The researchers found that it took the ‘truth’ about six times as long as ‘falsehood’ to reach 1,500 people. “We found”, the researchers concluded, “that falsehood diffused significantly faster, deeper and more broadly than the truth”.
A vivid example surfaced here in the UK this week of how a false claim can get half way around the world before the truth can get its pants on. The suggestion that the BBC TV Newsnight programme photoshopped a picture of Labour leader Jeremy Corbyn to make him look, as the Labour activist and journalist Owen Jones claimed, like a “Soviet stooge” had a social media reach of over two million, according to the Guido Fawkes website.
By contrast, tweets from the BBC debunking the untrue claims received just 2,400 retweets.
Jones alleged on Newsnight that “you had Jeremy Corbyn dressed up as a Soviet stooge, you even photoshopped his hat to look more Russian” – an assertion the BBC debunked. But the claim went viral. A video on Momentum’s Facebook page claiming the BBC “dressed up” Corbyn had 716,000 views and 20,000 shares. Momentum’s Twitter video had 266,000 views. This and similar videos had more than a quarter of a million views. By contrast, the BBC Press Office denial was retweeted just 600 times.
We are, it seems, more susceptible to ‘false news’ that reinforces our own prejudices than accounts of impartial truth. Responsibility, it seems, lies as much with the recipient as with the messenger – a point we should not overlook in the Cambridge Analytica scandal.