The Humanity of Computers

I love learning about how things work. And I loved today's discussion in my Digital Culture class about Meaning and Machinery, particularly the idea about machines as humans--or humans as machines. When one of my classmates said that Amazon knew she was pregnant before her mom knew, I was shocked. As I began to think about it though, I realized that many sites knew a lot about me: Facebook knows exactly how to advertise to me, Netflix recommends movies to me based on what it thinks my interests are, etc. How in the world does the computer know what I like--sometimes before even I know it?

This blog is going to discuss three things.
(1) How does the computer learn things about us?
(2) What is the controversy surrounding this learning?
(3) Machines as humans--or, humans as machines

How the Computer Learns

 

There is a buzzword that encapsulates how a computer learns about us: Big Data. Watch this short video to get a quick overview about big data and then I'll expound.

As the video introduced, Big Data is everything you do online. "Everything" meaning clicking a link, sharing information about yourself on a social networking website, what your friends "like" on Facebook, how long your mouse lingers over a picture, how often you check your email, or even what you say in that email.

Yes, the computer reads your emails. It reads your status updates. It recognizes faces of people in pictures because you once tagged yourself or others in those photos. It recognizes your voice because you uploaded a video of yourself speaking. It doesn't only know what you like. It knows who you are.

But how does it do all of that? Computers only work because of a bunch of zeros and ones, right? It should only be able to do whatever we directly tell it to do, right?

Right. And Wrong.

Computers learn things because they are told to learn things. They learn because they are told to store Big Data and then to infer things about people from that Big Data. We tell it to learn and to infer given certain parameters. Within those parameters, the computer then learns whatever it can. It does this through things as old as cookies to tracking users through screen resolution (no I'm not kidding).

Usually, your information is just sold to data brokers who then turn around and sell it to advertisers. But sometimes, things you'd rather not be exploited can be without your knowledge or permission.

 

The Privacy Debate

This leads us to talking about privacy debates. Should computers be allowed to do this? Well, they already are. Is there any way to stop them? And if not, should we be worried?

This XKCD comic summarizes the various views on internet privacy nicely:

http://xkcd.com/1269/

I'm a Sage. My husband, a computer engineer, is the Philospher/Crypto Nut. He likes to frequently bring up his suspicions regarding Big Data and I like to tease him about being paranoid.

Is he being paranoid? Or am I being too carefree?

88% of 865 Data Analysists (people that analyze Big Data for a living) think that "consumers should worry about privacy issues in the big data era, as more organizations stockpile personal--and often sensitive--information on all of us." Why? Well for one thing, there currently aren't any laws in the United States regulating Big Data. Some companies do put policies in place regarding Big Data that they collect (such as pharmaceutical companies), but they do not have to.

This has become a huge debate as far as the National Security Agency (NSA) and the public are concerned. There is something called the NSA PRISM. Author Al Mauer writes:

 "In the digital world of the 21^st century, all of our communications—other than face to face—are translated into a series of ones and zeroes, broken down into small groups called packets, disassembled,sent over a series of connected networks known collectively as 'the internet' and reassembled at the other end. Each of those little packets contains the routing information that allows the message to be reassembled and delivered . . . All of your communications go through gateways. . . Gmail, AT&T, AOL and many others . . .Using keyword searches, as Edward Snowden described, they can get to any information they have pretty quickly. They no longer have to capture and analyze the data as it comes in. They have it recorded and ready whenever they need it."

So...your medical history, your insurance information, your bank account locations, and more: the government now has access to all of this without a warrant which violates the 4th amendment to the Constitution. And in fact, the NSA has gathered this data without warrants from smartphones.

But, so what? If you have nothing to hide then you have nothing to fear, right? In fact, it'll help us catch terrorists! The end justifies the means.

But really, how many times has the NSA caught a terrorist through PRISM? None, as far as I can tell. This guy agrees. Therefore they are using it for other reasons. 

These are the two sides of the big debate surrounding internet privacy, particularly regarding the NSA. I won't go anymore into it now, but if you'd like, I can write a longer blog post analyzing both sides later.

Humans and Machines

We, as humans,  have created machines to collect Big Data, analyze it, infer things from it, and give us results. We have taught it to do this because we want it to be a tool for us to use. Yet, the machine is learning on its own.

Now, I do not think that the machines are going to rise up against us like in I, Robot or The Terminator. But there are some similarities between us. Here is a very simple mathematical data graph:

http://en.wikipedia.org/wiki/Graph_theory

There are three ways to interpret this graph:

1) As a Social Network. You are Circle 1. You are friends with Circles 2 and 5, who are also friends with each other. You are also connected to Circle 4 because Circle 5 is connected to Circle 4, even though you yourself are not connected to Circle 4. You are also connected to Circle 3. And Circle 6. All of these circles relate back to you because they are connected to you in some way.

2) As a Brain Network. The circle represent neurons. Neuron 1 fires off to neurons 5 and 2 which cna then fire off to 4 and 3. Neuron 4 can fire off to six. Neuron 4 can fire off to 3 and 5. Everything is connected to everything.

3) As a Neural Network. In other words, an artificial intelligence network (a computer's "brain"). Artificial Neuron 1 can fire off to artificial neuron 5 which can go to 4 and you get the point. This is a machine's intelligence. This is Big Data.

Machines think like humans. We learn from buzz words, collecting information on those buzzwords, analyzing, and then filling in the gaps. So do machines. You've heard that the brain is just like a big computer, right? We like to say that machines are inhuman because they lack common sense. Yet they don't. They infer things just like we do. I think I finally understand what the video "The Machine is Us/ing Us" from Dr. Burton's post is getting at.

Are humans essentially machines? Maybe. But I'm not so sure. Humans can feel emotions. Machines cannot. Humans can rebel. Machines do what they are told to do. Machines have elements of humanity and humans have elements of the machine, but we are still different. We are becoming more and more alike every day though.

Fascinating, isn't it?