
In the big story that exploded last week, and has been doing the rounds for a while. A data analytics / personal profiling / election campaign assistance company used personal and private data they scrapped from Facebook via a researchers quiz app. They then used this in to help influence the US election, and possibly the Brexit referendum. Facebook claim this is not a data breach, but a misuse of the old version of their platform api.
I’ve been wondering what to write up about this since the story broke, trouble is there is always new developments forthcoming.
To explain the core part of the story. In 2014 a researcher Aleksandr Kogan created one of those quiz apps, thisisyourdigitallife. The kind that used to annoy you on your timeline that your friends has taken. Due to the way Facebook worked back then, they where able to extract a lot of personal information about who was using it, and was able to do the same with that persons friends.
This information, according to reports, was later passed on the Cambridge Analytica who they used it to create psychographic profiles which was then used on behalf of their clients, particularly in elections, to target people with very specific influencing advertising, so called dark ads.
All of this came to light due to several whistleblowers, and a year long investigation by New York Times, The Guardian, Channel 4, and particularly the work of Carole Cadwalladr.
Theres a lot of ground to cover here, and so this is the light take as further posts may take a topic in depth.
- There is Facebook as a data mining platform and how its api was and what has since changed?
- Who these companies that bought the data are, what they did, and the question of it their techniques even work?
- How much influence do social media platforms have and what needs to change?
Facebook is a personal data mining platform, always has been. This goes back to the reported infamous chat log. Facebook and services like it exist to get you to put in as much personal information as possible. We may think of them as social spaces, public spaces event. They are not however, this is a walled garden and the cost of entry is your personal information. Aral Balkan describes it as surveillance capitalism.
Before 2015, Facebook would offer almost everything it has to any third part developer. I was surprised a lot by what I could take when I was dabbling with Facebook applications around 2010. By 2015 Facebook has locked it down quite a bit, unless your application asked for extended permissions. I remember working on a house sharing start up (Ruumi) when I was asked to build a feature to help someone who had signed up quickly inside the site invite friends to review them, and found that actually Facebook would only send a friends list containing others who had also signed into the site using Facebook.
Now since the researchers app was before the lock down happened, they where able to take all this personal data straight out of Facebook. It was only a terms of service limitation to stop passing that data around, it was all on a matter of trust.
If the phycological and big data targeting techniques claimed to have been used work it would be profound and even the ability to attempt this is very scary. That said, amongst my peers who are more focused on social media, there is a degree of scepticism on whether these techniques and tactics actually work, and if they had any influence. That the claims being made are the same as social advertising agencies that this all works, when really theres isn’t much evidence to really show that it does. I would be interested in seeing any metrics on this and how it was worked out if it converted any votes. Its also worth noting that Mark Zuckebergs took out print advertising to reach a vast audience to apologise on behalf of the company. Perhaps he didn’t fancy boosting his Facebook post to reach a wider audience.
I saw the #deletefacebook movement, but mostly I saw many friends saying they would not leave Facebook. That they valued the space and saw it as a way to connect, to promote and assist the causes they cared about, and to build a community. For my own use, I’m a user of Facebook, though its mostly posting in from elsewhere. Mind you, this elsewhere is also Instagram, another Facebook owned property. I did try to briefly promote (as I often try to) my presence on Mastodon, just not seen any takers. Facebook does provide something exceptionally useful to a lot of people. However given how abstract the concept of taking personal data seemed, many of us took it with a pinch of salt, after all what could possibly be interesting about my life to advertisers. I have my sympathies with why people find it hard to leave, given that ‘everyone is on Facebook’. Though for many of us, we have been alive longer than Facebook has existed, and got along and formed social groups and promoted causes and did good things fine.
I do think the issue raised with this, the mining of personal data to the influence elections, and the other issues (Russian Government bots) are important to look at and build defences against. I do worry this means we loose sight of some of the deep divisions that have happened in our societies. Its important not to fall into the trap of just blaming the tech, its misuse, and bad actors, and if it wasn’t for that then it would all be ok. The campaigns these techniques are said to have been used on where already marked as competitive, with a lot of anger from left behind people. I think rebuilding social and community relationships, perhaps in an offline first way and including with groups that voted ‘the other way’ (excl nazis obviously) is going to help us move forward than just fixing the tech alone.
Footnote : Took me multiple times to spell Analytica correctly, and Apple spellcheck still insists it’s wrong.