Artificial Intelligence and Machine Learning, or AI and ML for short respectively, are two terms which get thrown around quite often these days. Most people think that they share the same meaning, and while they are quite closely associated, one cannot be used instead of the other. Both of them crop up frequently when discussing analytics, data or any sort of technological change, so I think it’s important we settle what they actually mean once and for all.
In short, artificial intelligence is the general concept where machines are able to carry out tasks in a “smart” way. Or, at least, in a way which we would consider smart. Your microwave is not a smart machine, very much like your fridge. It’s an appliance made to carry out one task and operate in the same way every single time. There’s nothing complex about its operation.
AI has been around for an extremely long time. Old Greek myths suggest that mechanical men designed to mimic our behavior have been around for over a century. Even the earliest European computers were considered “smart” and “logical”, despite the fact that they could only do basic arithmetic calculations and had short memory. The simple fact of the matter is that their creators designed them to, in a way, recreate the human brain. Create a sort-of mechanical brain if you will.
As time progressed and machines got more and more intelligent however, our concept of what counts as a “smart” machine changed rapidly. Nowadays, even your smart TV can carry out tasks far more complex than anyone could have imagined just ten years ago. Artificial Intelligence is taking a new direction these days however, and it’s far more complicated and difficult to pull off than anything we’ve attempted before.
Rather than trying to give machines more and more complex tasks, AI’s goal is to mimic human decision-making and carry out tasks and jobs in human-like ways. Artificial Intelligences (devices which act intelligently) are divided into two main subcategories: applied and general.
Applied AI is extremely common these days. An applied AI devices is any system which can carry out simpler tasks such as trade stocks and shares, or even maneuver a car through an autonomous system. You get the basic idea. Smart devices and/or systems which are primarily designed to carry out one thing, and one thing alone. They’re really good at the particular thing they were designed for, but can’t operate beyond that scope.
Generalized AIs can in theory handle virtually any task. They are not as common as applied AIs, but it’s why they’re so exciting and interesting. Most of the progress and innovation comes exactly from generalized AIs. Generalized AI is also the birthplace of our next talking point, Machine Learning (ML).
Machine learning was founded and based on two very important breakthroughs. The first came back in 1959, when Arthur Samuel realized that rather than teaching computers everything they need to know about the world, it might be possible to teach them how to learn for themselves. Much like us humans do. Our parents don’t teach us everything they know or everything they think we need to know. With time and through experiences we gather our own insights into how the world works, and shape our own opinions. Why can’t the same hold true for computers?
The second realization came with the emergence of the internet. Back when the concept of ML first came to be, it was just that, a concept. Even if you could, in theory, program a computer to learn all on its own, it had nothing to learn from. The internet gives you full access to practically all of the information in the world. It’s the perfect learning source for a human-like AI, and the perfect proving ground which will hopefully show that ML really can work.
Neural networks proved to be vital in teaching computers how to think and operate. In short, a neural network is a computer system which has been designed to work in the same way a human brain works. It can recognize, classify and sort all sorts of images and text, very much like you and me. It still holds advantages over the human brain though, such as speed, accuracy, and lack of any bias.
The end goal, for me at least, is natural communication between humans and AIs. You should be able to talk to a computer or a device in the same manner you would talk to another living being. That is, after all, the sole purpose of AI and ML, to recreate the human brain and its way of thinking into something which is not inherently human. This can greatly benefit humanity, but it also raises a couple of moral questions.
For starters, what happens when computers inevitably reach human-levels of thinking and consciousness? I’m not talking about D-Day or computers taking over the entire world, but rather the moral point of view. Do we lose humanity as a trait when that happens? Life’s greatest gift is that you’re alive and can experience everything around you in your own, unique way. If a machine is capable of doing the same, I think it kind of degrades the value of life a little bit.
It’s like finding out the meaning of life. If something as simple as a couple of circuits, screws and wires can recreate what you have as a fully living organism, doesn’t it make life less valuable? These are all questions we’re going to find answers to relatively soon.