top of page

The Black Box Problem

“One would think that by the second decade of the twenty-first century, the intellectual poverty of technocracy and the primacy of politics over it would be a well-established truth in need of no further defense.”

-Evgeny Morozov

Currently, I'm sitting on a patio in Florida, metling not only from the heat, but also from the intense mathematical equations written on the paper in front of me. They involve Bayesian Statistics, Kalman Filtering, and State Space Equations. This however, is of little importance. After researching those topics for a little while, I decided to take a break. I opened up my soundcloud, admired the lake surrounding the patio, and then decided (very poor judgement, by the way) to open my Facebook. I scrolled down and then happened upon this article: http://www.theverge.com/2017/3/30/15124466/ai-photo-style-transfer-deep-neural-nets-adobe​

I'm positively astounded. There is now the ability to transfer artistic styles between different images. The clear blue sky above me cracked open, and the angel of death descended towards me. Perhaps it was the dosage of intense math and incredible technological development, but I had a sudden "oh boy" moment. I'm thinking that we're not headed for a very friendly place in the future. Put simply, I'm thinking that the intense pace of technological development will make it incredibly difficult for any person to keep up with the latest research in order to defend themselves from what I like to call the black box problem of authority, which I originally read about in The Seventh Sense by Joshua Cooper Ramo. It's an excellent book. I would highly recommend it to those that want to ponder further the interaction between political systems and technology. But I digress. I think that as the technology around us becomes more advanced, we will become so ignorant of how it works that those that develop the technology will begin to have complete control over their users. And when I say complete control, I mean it. Through the use of metadata and data mining, companies will (and do) know more about our lives than our family or significant others do. With this knowledge comes the ability to predict many if not all of our future desires and our actions. However, this lack of knowledge about the systems we use is only an issue if we are reliant upon the system. Are we? Look around you. Most everyone has a smartphone, and most everyone is nearly addicted, especially among my generation. In a way, even though we're all extremely familiar with technology like a laptop, smartphone, facebook, and so on, only a few people, the technocratic elite, actually understand the ins and outs of all the components of those systems. They are almost all powerful, because they have such a wide consumer base. Remember when Facebook ran social experiments on its users?: https://www.wsj.com/articles/furor-erupts-over-facebook-experiment-on-users-1404085840. Nothing is to prevent them or any company from doing something like that again, indeed many companies are. They already sell your data[1], and with the new bill to let Internet Service Providers directly sell search history, this is now certain to be extended even further[2]. This is a problem of abstraction, of never really being certain of the output of a system since you don't know what's inside it, like any black box method or algorithm. Any member of the new elite can modify what's inside the black box without our knowledge, and our lives could be subtly altered in ways we don't understand because we don't know what's inside. It can then be said that with the large degree of reliance on our almost wizardry like technology that we no longer control a large portion of our lives. This is undesirable, for me at least. I always like to possess as much latitude in controlling my life as possible. I suspect most of us are the same way.

In my mind, there are three solutions to abstraction's grave threat to a functioning, powerful, knowledgeable middle-class and individual freedoms. The first is to educate yourself relentlessly, never stopping in your quest to understand at a basic level almost every single technological development. This is extremely difficult, verging on impossible, for obvious reasons. The second option is to distribute technical knowledge to a large degree, preventing any one person from knowing the entire puzzle. This seems feasible, since specialization is already a driving force of the American market economy. The third option which is very robust, but also has its tradeoffs is to stop personal reliance on these developments if you do not know how they function. Simply put, only use things that you can understand. Unfortunately, this seems to be a long shot since everything the modern American consumer uses has so many levels of complexity that have been abstracted away by a few buttons or maybe even no buttons at all. None are perfect solutions but I think developing distributed frameworks for the development of new knowledge is the most realistic option. I'd be curious to see what you guys think.


bottom of page