Facebook as Google for Criminals
Peter Burnett, author of How to Do Privacy in the 21st Century, argues we should be surprised at the surprise over Facebook and Cambridge Analytica.
One of my preferred places on Facebook is the Bella Caledonia Public Group. It isn’t the echo chamber some might imagine – there are vastly differing viewpoints across the left and it’s possible to have a sound discussion in this group with people from all over. Many visit this Public Group just to see what’s being said, as a form of news gathering, or to widen perspective or sometimes even to narrow it. Some go there for a fight, and there are always a few of these too.
Looked at from the outside however, even the Public Group is a potentially dangerous prospect. What better place to gather in a matter of seconds the names, preferences, images, friends and families of thousands of people whom some might wish to broadly brand as YES supporters, socialists, egalitarians, anarchists, republicans, communal thinkers and other believers in self-rule?
As recently as two decades ago, political canvassers and marketeers of all stamps had to guess how individuals and thus the collective would behave in elections and other situations. The problem at the heart of the Cambridge Analytica news this week, is that since this information is now readily available, the challenge is as Analytica puts it themselves, “to use data to change audience behaviour”.
Why is Facebook highlighted so, however? Every large company collects data, and every small one does too. My own modest small business collects the email addresses, ages and gender of its customers in order to contact them in the hope they’ll use our service again. We are all gathering data, but the issue with Facebook in the Cambridge Analytica case is not the scale. The issue is that active collusion between political parties and these firms is frustratingly easy to chart, but hard to prove.
As a Facebook user, you are not just providing data, but you are collecting it also. Facebook want you to do this and that is a part of their power and charm. If you are in business, the plan is simple: when you make money, they make money.
That is to say, if you are selling a product or service using Facebook, the more of it you sell, the better they will do too. It’s a corruption of the capitalist model whereby we become the product. Imagine your data as your surplus if you like, and the more of it you have, the more value you have as a product.
As a political product, your discomfort may increase, and we must all wonder about our chances in any forthcoming referenda and elections. The first US presidential election to involve fake news was in 1844. One story involved a newspaper report about an alleged slaveholder who apparently met with some of Democrat James Polk’s slaves cruelly branded with Polk’s initials, proof it was said that Polk had sold slaves to raise money for his campaign. It was not true.
In the UK, Joseph Ball, the Conservative Party’s director of publicity placed a spy in the Labour Party’s printing works, in the 1930s, and in the absence of Facebook, that was as bad as it could get.
The only wisdom which can here be derived is that we should work harder to vote with our heads, and not our hearts, which is easier said than done, especially given that Cambridge Analytica have been captured on tape stating that it was they who made Trump a reality. They are not the only people boasting about this, I should add. Roger Stone takes credit for Trump’s win in much the same way, and even Nigel Farage wants to be associated with this piggish victory, arguing that it was Ukip and Brexit which “directly led” to Donald Trump’s presidency.
Once more, we should be surprised at the surprise. The Guardian video of Christopher Wylie, the so-called whistle-blower in the Cambridge Analytica case explains nothing that we do not know already.
Shock! “They had apps on Facebook that were given special permission to harvest data not just from the person who used the app, or joins the app, but also it would then go into their entire friend network and pull out all of the friends’ data as well.”
The misinformation in this sentence alone (published in The Guardian) is not helpful. The permission in this instance was not ‘special’ insofar as it pertains to any Facebook app; and it could not pull ‘all’ the data of a user’s friends. What is described is of course carried out by tens of thousands of companies and apps, and done so all the time because it is what Facebook was built for. The most misleading part of this interview is the question:
“And people had no idea their data was been taken in this way?” which query returns the answer “No.”
I call this misleading because each responsible user of Facebook assumes that they are not subject to privacy. There are settings which may minimise this, but it isn’t much of a jump to assume that advertisers and campaigners are digging deep here and selling us shoes, snacks, presidents and withdrawals from economic unions that we don’t need. Worse is the claim the claim that Cambridge Analytica spent in Christopher Wylie’s words, “a million dollars” on this. It makes them sound like Dr Evil, from Austin Powers.
“It looks today as if psychology researcher Aleksandr Kogan is going to be hung out to dry by both Facebook and Cambridge Analytica as a momentary distraction while we go back to those videos of people and dogs sliding about on ice. All he wanted to do, argues the psychologist, is predict how people will behave. What is sure is that business will continue as usual with Facebook users learning nothing they did not already know.”
I highlight this, because as you are aware, the data is freely available and a million dollars is not required to access it. A fellow contributor of mine at the magazine 2600, Anthony Russel, demonstrated this last year, in a proposal he titled Converting the Voter Database and Facebook into a Google for Criminals.
Using freely available info, Russell made a proof of concept app that took the Ohio voter database and linked it to Facebook and found out that he could be accurate 45% of the time. The result of this was that if the voter database he used contained 6.5 million people, he could be confident about extrapolating 2.86 million records by matching them with Facebook. The process was ultimately simple, and it began with the public electoral roll, which contains the first name, last name, date of birth and home address of everybody above age.
“To start,” said Anthony, “I had to see what public data was available. In short, there’s a ton. No wonder we get marketed nonstop by mail. The government takes our personal information and puts it on the web for free. Write a couple of scripts and you can tap it anytime.”
To begin with then, Russell accessed public voter records, which contained the above details of 6.5 million people in a straightforward CSV (Comma Separated Value) file which could be opened by any spreadsheet program.
What he did then, was to make the same basic Facebook queries to find people in a certain area, and when he had them, he continued hit these records with a search which would have probably looked something like this:
Once he had generated a list of potential matching profiles, my colleague generated confidence scores, to predict the accuracy of any given result. Using the names, dates of birth and addresses that he had accessed publicly, he could then synchronise this data and present a confidence score that this profile match was accurate.
Russel presented his redacted script in 2600 magazine, and wrote: “I was able to run this script over thousands of people without getting rate limited by Facebook. Conceivably, I could run this nonstop and eventually build a giant database.”
Such a process could leverage all sorts of data which we might imagine is private, but which is commonly attainable. It could offer all kinds of information, such as all the people that work for a specific company, or all the elderly people in a certain area, information on groups, families and public services, including their personnel. Such a database could also target people by interest, and be used for marketing, or indeed any of many criminal abuses.
Russell concluded: “Allowing anyone to download the entire voter database is probably a dumb idea. I understand why voter records are public and it’s for a good reason. That said, the government just enabled me to build a Google for criminal enterprises. Facebook should also probably be limiting the above queries.”
It does appear here that Facebook data has been used for criminal ends, as it is not legal to influence elections by spending such large sums on advertising which pretends not to be affiliated with a particular candidate or party. There are some rules of the jungle that are being followed very closely at the same time, though. Take this statement from Lynton Crosby:
“At its absolute simplest, a campaign is simply finding out who will decide the outcome … where are they, what matters to them, and how do you reach them?”
Crosby, who ran the Tories’ 2015 UK General Election campaign and who also saw Boris Johnson elected mayor of a largely left-leaning London, has been preaching this gospel since the 1990s, long before Facebook began, and he wins elections because he is skilled at voter-targeting. And I would imagine that in any election, voter-targeting plays a significant part, even though ethical questions will abound when people who will do anything they can to win, are employed.
It looks today as if psychology researcher Aleksandr Kogan is going to be hung out to dry by both Facebook and Cambridge Analytica as a momentary distraction while we go back to those videos of people and dogs sliding about on ice. All he wanted to do, argues the psychologist, is predict how people will behave. What is sure is that business will continue as usual with Facebook users learning nothing they did not already know. We can agree with Cambridge Analytica’s executives insofar as elections are on the whole about emotions and not facts.
But nobody, and I mean nobody has quit Facebook in despair and outrage at this story. We can therefore hope that, just as with the LABOUR ISN’T WORKING poster of 1979, we don’t continue to vote on the basis of memes, but on issues, because that is exactly what has happened here.