MIT Technology recently reported that Petko Bogdanov and colleagues have developed a genetically-inspired approach to predicting a user’s social media usage. The technique identifies a user’s social media ‘genotype’ (a mapping of what the user is interested in and what he or she shares via social media) and uses this genotype to predict what a user will be interested in looking at, learning about or purchasing in the future.
As fascinating as this information-gathering technique is, make no mistake, this is dangerous. Lately, everyone is concerned with privacy of information, especially with the NSA wiretapping watershed. The identification of a person’s ‘genotype’ can make information gathering, advertising and information distribution much easier for individuals, companies and agencies. While the technique and nomenclature that Bogdanov, et al. have developed are unique, the ideology behind them is nothing new.
For years, the internet and the way websites and networked technologies function has been moving towards a model of giving people what they want the moment they want it. Your Amazon, Twitter, Google, Facebook and other accounts all share information with each other and with advertisers. This information follows you everywhere, not tied to one device. We access all this through our computers at home and at work, our laptops, tablets and our cell phones. Furthermore, we are storing all this information in the cloud, which can be accessed from anywhere by the user, but it also puts a lot of information into the control of a lot of people you don’t know – the NSA being one scary example. And this information can be accessed without having to access a specific physical device. On one hand, this technology networking allows you to see the information you want very quickly and easily. But on the other hand, it lets advertisers and other agencies easily track you and find out a lot about you. It’s a trade-off and should be used with caution. Yes, these services and technologies are great and can help you find a lot of things that you’re interested in without wading through a lot of stuff that you don’t care about, but you have to take into account how much that’s worth to you. There are hidden costs to “free” services like Amazon and Google and as information becomes more valuable and is being traded around the information industry like currency, we have to be careful not to be throwing our information around like we’re buying penny candy.
Besides the danger of giving up information you may want to keep private, there is a much more intellectual danger involved in the personalization of your internet experience. As the services a user accesses adapt to a user’s ‘genotype,’ they start to see more of the same information – information that specializes and pigeonholes their knowledge base and reaffirms what they already believe.
Let’s use myself as an example for exploring someone’s genotype and how that may affect their experience. First of all, I am not a great “node” for people to look at, as the first thing my genotype would say about me is that I am not very likely to propagate information via social media and any of the few hashtags I use (a big factor in determining a person’s genotype) are likely to be very sarcastic (i.e. #YOLO…yeah, because I really buy into YOLO). But anyway, my social media genotype would be broken down into a few categories, with subcategories underneath them which determine my specific interests and genotype.
- Business: Not much, except when it relates to technology, and that rarely
- Celebrities: Not much interest, but I do follow a few comedians on the Twitters
- Politics: IP & copyright policy, digital rights and activism
- Sci/Tech: Privacy, networked technology, space exploration
- Sports: Hockey (Chicago Blackhawks and Philadelphia Flyers), Basketball (mostly college but also Chicago Bulls), Football (again mostly college but also Chicago Bears), Tennis (Men’s and Women’s singles)
It is also interesting to think of this genotyping technique in terms of forming collective intelligence, specifically collective cultural intelligence. In the introduction, the authors state:
The social media genotype, similar to a biological genotype, captures unique user traits and variations in different genes (topics). Within the genotype model, a node becomes an individual represented by a set of unique invariant properties.
Each individual functions as an unvarying node in a giant web of culture and information sharing. If each of these nodes can be identified and predicted accurately, as the authors claim, it is possible to look at this web as a system of inputs and outputs. If one user inputs a certain set of data, its travel through the web of nodes (users) can be predicted and we should be able to roughly determine how far the data will go through the web and how many users it will reach. This is not just a powerful tool for advertisers and other businesses, but it could have a great impact on cultural studies as well. Memes, art, film, music, everything becomes trackable and predictable.
The difficulty lies, of course, in the level of unpredictability in the users. Sure, I am likely to follow or have interest in certain sports teams, video games, types of art and music, etc. but it doesn’t mean I will. And I don’t see everything that comes through the web related to those categories. Sometimes I don’t access social media or news via the internet at all. Does that just make me an unreliable node? The authors do have a lot of algorithms to determine the success rate of understanding a user and predicting how useful that information is. See this table from the article:
To what extent can we rely on this type of technique? The authors admit to inconsistencies among individual users and group users together, but can we get even better at predicting the usage of individual users, thus turning them into nodes? How valuable is this information on someone who doesn’t participate as fully as others in social media. And what will the ramifications of this type of technique be? How will this change the way the internet functions and the way we use it?
One thing is for sure, though. A step closer to predicting and determining cultural exchange on the internet is a step closer to forming a better understanding of networked intelligence and networked culture. Will this lead to the rise of an omniscient machine intelligence? As cool as that would be, we’re probably a long way off. But this does lend credence to my prediction that should we create artificial intelligence, it should be able to understand culture as well as pure information. This is because as we network our interests, our art, our cultural products with each other, we are not just disseminating information but ideas as well. Perhaps we can give the first true artificial intelligence a social media genotype. I vote Chicago sports, stoner music, American modernism and radical digital liberalism.