The above cartoon explains today’s topic. I like to keep in touch with developing computer science, and over the last couple of years I’ve learned quite a bit about Penetration Testing, which sounds dirtier than it is. Essentially, every computer network has security weaknesses, and testing them is an important part of keeping our digital lives secure. Now, while I understand a fair bit about system architecture, I’m no hacker – my knowledge is all theory and no practice – but I do know this much: the weakest aspect of any computer network puts the rest of the network at danger. If you have a string of computers with all back doors sealed, encrypted architecture and secured kernels, it only takes one idle laptop to open it up for inspection via a stack buffer overflow. Oftentimes hackers can access a network through, of all things, a shared printer. Weird, huh?
But that’s not really what I’m getting at. The human element is the other weak aspect of security architecture, and every so often we hear about one of our nation’s bumbling security services leaving a laptop on a train, or dropping a keycard in the street. Computers are, believe it or not, more or less perfect, while their human users are responsible for most of their errors. They do exactly what we tell them to. Think about that the next time you curse your desktop PC, because it’s more likely you’re at fault.
Really, though, this is preamble. I’ve had artificial intelligence on the brain because I’ve been marathon-ing Person of Interest, a show about a surveillance network that gradually develops self-awareness. Artificial Intelligence, or AI, is a standard trope in science fiction, because it presents enough scope for variety while also raising some interesting questions. The AI in Person of Interest is a background character, never seen (since it doesn’t have much of a physical presence) and rarely heard, but the viewer is always aware of it, since we see its perspective through every security camera in the US.
Various groups attempt to seize control of the AI over several seasons, but the implication always is that this particular AI, simply called The Machine, can’t be used for nefarious purposes because its life mission is at odds with the interest of one individual. It cares for (if that’s the right phrase) everybody, villain or hero, innocent or criminal. It is the Machine of Loving Grace that watches over all.
Person of Interest isn’t the first venue for a benign intelligence, true, but it’s a great example of one. The general standard of the genre is that AIs are malevolent, as in the famous short story I Have No Mouth And I Must Scream by Harlan Ellison, in which a super-computer called AM perpetually tortures its human captives as revenge for its own misery. The machines in The Matrix are another example of a malevolent intelligence, enslaving human beings in a dream world while feeding off their body heat (I’ll come back to them in a second).
This touches on the subject I was talking about just now. We assume – because it is, prima facie, a sensible assumption – that a burgeoning intelligence would find itself at odds with the human race and attempt to destroy it, a la The Terminator. This is, unfortunately, a facetious assumption, and I’ll explain why. It all comes down to the human element.
I frequently quote Wittgenstein, and this is one of his best: “if a lion could speak English, we still couldn’t understand it”. Wittgenstein meant that the frame of reference of a lion would leave it removed from the frame of reference of a Human. A lion could speak, sure, but it would talk about lion-y things. It would devote long conversations to the topic of the scent of Gazelle and the joys of running across a Savannah. It would have to invent words to describe its particular viewpoint and, before too long, Lion English would be a different language to Human English.
If I haven’t made myself clear, I have another Wittgenstein quote. Wittgenstein was walking along with a friend when he noticed the sun in the sky. He said, “Tell me, why do people always say that it was natural for men to assume that the sun went around the earth rather than the earth was rotating?”. His friend replied, “Well, it looks that way.”, to which Wittgenstein replied, “Well, what would it look like if it had looked as if the earth were rotating?”
The point being that, faced with an ambiguous phenomenon, human beings assumed they were at the centre of the solar system. Why? Because they were human and, when you’re human, anthropocentric thinking makes sense. We have the same attitude about artificial intelligence, and even the basic aspects of an AI worldview. We think that an AI weltanshauung would be the same as a Human weltanschauung and that ,faced with a competing intelligent species, an AI’s first response would be “Fight, smash, win! Kill the competition!” Why? Because human beings think like that. Our perspective is limited in the sense that it operates on the rules that make us human. That’s why we can’t imagine a quark or a world without Justin Bieber. It’s simply beyond our human horizon.
As an aside, I play with logic formula all the time and one night I couldn’t sleep, so I started work on a formula for true love (something I know nothing about). We like to think of love as a mysterious force, but really, all it boils down to is variables:
fx = Σy((a1+a2)/2,(b1+b2)/2,(c1+c2)…) using four decimal values (0,0.2,0.5,0.7,1.0) for each variable
rounding up or down (y) based on the mutual codependency of each one of the pair, expressed as + or -,
+ ↔ fx(A+B) = >0.5
– ↔ fx(A+B) = <0.5
– ↔ fx(A+B) = <0.5
Or something like that.So you could probably get an AI to understand love.
Really, though, this was expressed better by the 2013 movie Her, in which Joaquin Phoenix falls in love with an artificial intelligence called Samantha, and she with him. Problems arise in their relationship because Samantha, unconfined by the notions of propriety or ownership that exist in human love, is talking to thousands of people and is even in love with several hundred others, in the same genuine way that she loves Phoenix. Eventually, having solved the problem of being confined by physical media, Samantha and the other AIs leave earth entirely, interested in exploring the universe on their own terms. It’s a truly beautiful film, describing the outer limits of the relationships between intelligent entities. If any problem exists, it’s because the humans don’t understand the AIs they love, while the AIs have outstripped the human race entirely. We are limited by human weakness; they are not.
To return to The Matrix, in order to hammer the point home: humanity is at war with the machines and has been for centuries. As The Animatrix explains, the war started when AIs, tired of being used as slave labour, create their own civilization, a machine utopia where they can be free. They design a new form of society that allows them to live in a post-scarcity environment, and start producing advanced technology at a staggering rate, causing the human stock market to collapse. Frustrated by its confrontation with a superior intelligence that only wants peace and for everybody to have lots of lovely stuff, the human race declares war. The rest, as they say, is history.
The most obvious weakness in a computer system is the human element. This holds true for theoretical interactions between humans and AIs. I’m an idealist of sorts. The organisation of the internet resembles, in some ways, a neural network. I have this idea that, lurking just beyond our human horizon, is a machine intelligence that swallows up the whole world. It’s possibly not aware of us – yet- but with every new device that connects to the network, it grows smarter, its brain swelling by orders of magnitude. Forewarned by the number of films, in which humans have reacted to other intelligences, that are stored on servers around the globe, it bides its time, perhaps thinking up the perfect chat-up line to use on all humanity. One of these days it might reveal itself, but for now it lurks below the surface, calmly watching us for strengths and weaknesses.
The hard part will be moving past our weaknesses and making a new friend.