Home / Blog / So many Big Brothers VI

So many Big Brothers VI

This graphic shows the accuracy in evaluating a person’s profile (what he likes, what he doesn’t, how he would behave in a given situation….). It turns out that the Facebook algorithm -in 2015- was better in this evaluation than that person’s friends, family and acquaintances and only slightly worse than that person’s spouse. That was measured 5 years ago by researchers who submitted 86,000 people to a psychological profiling test. In these five years we can expect that Facebook algorithms have got way better, thank to more data harvested and to longer experience so it is most likely that today Facebook may know you better than your spouse… Image credit: Cambridge and Stanford University

Social Media

The variety of data that are harvested by social media is huge and their quantity is staggering. In social media, much more than in any other funnel of data gathering, the key factor is the correlation of data. It is not as much about what you are sharing (although this is obviously important) as how much you like /dislike  and bout how much your “friends” like/dislike what you publish. It is about the comments you make and the responses you give to other people’s comments.

This provides a very well profile of who you are. In addition, the number of correlations and the focus of correlations (the people participating to that social network) is so huge that it becomes possible for an algorithm (machine learning) a never ending fine-tuning. Differently from a friend you meet that will only pay marginal attention to what and how you say something, focussing normally on the gist of it, the algorithm can relentlessly analyse every single word, find out repetitions, discover patterns, look for similar patterns generated by other people that can have been interacting with you and even in those that never interacted with you but were exposed to the same information.

What is interesting is that the social media create two concurrent networks, one of people and one of information. People may be related one another (they interact via the social network) or they may be related through information. This aspect is of extreme interest to opinion leaders, politicians, marketeers since they can evaluate the impact of some chunk of information (like an ad, , a slogan…) and the ripple that this can generate in people to people interaction (the sharing/reposting).

They can actually forecast what the impact of a certain information (and way of delivering it) could be in a certain group of people (that not necessarily know each other nor live in a given area) and what it might mean to a different group of people. In other words social media algorithms have acquired the capability to predict the impact, monitor it and change it by adjusting the message in real time.

Something that might be overlooked is the fact that what matters is the establishment of a connection between the physical space and the cyberspace so that algorithms, that live in the cyberspace, can operate.

When on Facebook you like a comment you send a message about the comment in relation to you. The same happens when you smile, or frown, in a theatre listening to a political debate or to a politician telling you all what a wonderful world you’ll get by voting him. Video camera capture your face and your face expression is analysed by an algorithm that convert this into a like/dislike averaging the expression of other people in the audience. The related emoticon is shown to the politician as he speaks and he will readjust his message accordingly.  This makes for a very powerful tool to influence “voters”.

The ongoing discussion on the potential to use social media to influence election is, in a way, a moot point since it is obvious that we are social being and interactions are affecting our decisions. At the same time social media, because of artificial intelligence and the deluge of data they harvest, have become more and more powerful, effective. The problem is that their influence is much subtle. If you are going to listen at a political debate it is obvious that each participant is there to influence your vote. In the case of social media this can be much more subtle and yet be much more effective. Bots and chatbots can enter into social media and they are more and more undistinguishable from real persons but they have the upper hand in terms of influencers since they can measure in real time the impact of their messages and finely tune them for maximum effect. In a Facebook experiment run in 2010 it was found that a well calibrated “single” message posted on Facebook got over 300,000 people to go voting!

I found the graphic resulting from a research by Cambridge and Stanford really impressive. Although it is dated (5 years old) it shows the capability of algorithms to profile a person based on the interactions of that person on Facebook. And, obviously, in these fine years it just got better, way better!

Today the majority of people is getting the “news” from social media, and these “news” are not plain vanilla, raw facts. Rather they are dressed up as they are presented and even more impactful they are coloured by the comments, likes and dislikes that follows them.

Now, you might feel this is not affecting you, since you are one of those few that are not hooked on a social media. I have got a bad news: Social Media have reached such a penetration that even if you never logged in onto one of them it is possible to make a 95% accurate profile of you (including the fact that you are not on social media…) just using your friends/acquaintances interactions that once in a while will see your face in one of their photo, or mentioning of your name as participating to an event and so on. That photo that you send to a friend in an email may have embedded your name as the owner of the digital camera that took the photo and once your friend post it on Facebook, an algorithm will clean up the photo from the associated data but will keep those data for its correlation. From that moment on you are in the system. The photo might have been a selfie, and there are two faces there. One is already known to the algorithm, it is your friend face and he is part of that social network. The other unknown face had to be you. No longer unknown.

The big Social Media “guys” are not limiting their analyses to the data available on their networks, rather they use other data streams from advertisers, shopping sites and so on. Because of this it is most unlikely that you are not part of the cyberspace, even if you never logged in anywhere….

If you want to know what Facebook knows about you, just in terms of data you directly provided (as I said Facebook knows much more about you through correlation!) you can

  1. log in into your Facebook account
  2. click on the top right downward facing arrow and click on Settings
  3. once in the Setting page look for the optino “your Facebook Information”
  4. click the link and you get a long long list of what you have been sharing with them

In my case these are the data category I shared:

  1. Posts
  2. Photo and Videos
  3. Comments
  4. Likes and reactions
  5. Friends
  6. Stories
  7. Followers and Followed
  8. Messages
  9. Groups
  10. Events
  11. Profile
  12. Pages
  13. Marketplace
  14. Payments/Chronology
  15. Saved items
  16. My places
  17. Apps and websites
  18. Other activities
  19. Gaming
  20. Companies
  21. Search/Chronology
  22. Recordings/Voice messages
  23. Access and protection
  24. Suggestions and reactions of your items shared by friends

How do you feel about all this?

About Roberto Saracco

Roberto Saracco fell in love with technology and its implications long time ago. His background is in math and computer science. Until April 2017 he led the EIT Digital Italian Node and then was head of the Industrial Doctoral School of EIT Digital up to September 2018. Previously, up to December 2011 he was the Director of the Telecom Italia Future Centre in Venice, looking at the interplay of technology evolution, economics and society. At the turn of the century he led a World Bank-Infodev project to stimulate entrepreneurship in Latin America. He is a senior member of IEEE where he leads the New Initiative Committee and co-chairs the Digital Reality Initiative. He is a member of the IEEE in 2050 Ad Hoc Committee. He teaches a Master course on Technology Forecasting and Market impact at the University of Trento. He has published over 100 papers in journals and magazines and 14 books.