Artificial Intelligence for non-tech people
The tech world is buzzing with new discoveries in areas and fields that makes what humans have imagined for years plausible. The fact is that the reality is going to hit the world (as it has already) whether you know about it or not. Therefore it is important to be in the know. If it is just to be able to have something to say where conversations about these things are held, it is enough reason to learn.
Artificial Intelligence is a system whereby a computer or algorithm is designed to make decisions that are traditionally peculiar to only humans (or animals) on its own. This is enabling machines and computer systems to be able to make decisions we have previously believed only humans can. The core of it is that those systems can be able to make these decisions faster and more accurately than (most, if not all) humans. It is called Artificial Intelligence because it is intelligence outside the natural order that humans know and have been used to. Since the qualification for most of the work people do today is intelligence, a computer system that can do the same with better speed and accuracy is naturally preferred. This is why AI creates the debate for the future of work (for humans).
AI is not so much a future thing as it already exists today. There are various applications of AI that are functioning already. A simple example is the speech recognition software. Have you ever thought of how a computer system could recognize your voice? It is previously thought that only humans and animals can do that. But lots of computer systems do today. Another common application of AI today is in virtual personal assistants; (Apple’s) Siri, (Android’s) Google Now, (Microsoft’s) Cortana. If you use the iPhone, you must have been already familiar with Siri. These virtual assistants learn about the user to the point where they can anticipate the users needs. That is AI. There are other applications already existing now like video game characters, self-driving cars, online customer support, music and movie recommendation services, smart home devices, etc.
Another buzzword surrounding the AI initiative is Big Data. Big Data is not just data that is big (although that kind of puts it so simply). Big Data is data that is beyond what the usual data processing software and analysis can deal with. It is data that is so voluminous that can be analyzed computationally to reveal patterns, trends, behaviors especially those that have to do with people. Big Data is often what is fed into AI systems for them to be the efficient decision makers that they are.
There are key subsets of AI that ought to be understood also. The first one is Machine Learning. Machine Learning is inside AI, but there are aspects of AI that are not Machine Learning. The common aspect of AI that is not in Machine Learning is often referred to as rule based. This is when the computer is designed to follow a kind of set of instructions to make intelligent decisions. This is quickly fading out because of a much stronger model which is Machine Learning. Machine Learning (as the name implies) is simply when the system is taught how to learn and mimic (the desired) human behavior. Machine Learning is when the system learns on its own by data (mostly, big data) without any programming. It’s just like you feed the system data, the system analyzes it to find patterns, trends, etc., then it makes takes decision based on the task assigned to it from what it has gathered or learnt from the data. A simple example of this is a computer game test demonstrated by some students in Europe. They fed all information required to note (i.e. what is not allowed) before playing the game into the AI system and gave the system the objective of the game. Without having any information about playing the game or prior knowledge about how the game is actually played, the system played it in the most efficient way. This is Machine Learning.
Machine Learning is where AI begins to get scary for a lot of people. And it is not supposed to be, the problem is that we have fed our minds with all kinds of terrible ideas through movies, novels, etc. Most AI today uses Machine Learning. Although the decision making of the system (in Machine Learning) is not a result of direct programming by any engineer, there is still need for an engineer or programmer to certify the result or decision (by the system) as correct or incorrect. Machine Learning still requires supervision. But there is a subset of Machine Learning that tackles that. This is where it gets crazier.
This next buzzword is called Deep Learning. Remember Machine Learning is inside AI, now Deep Learning is inside Machine Learning. Deep Learning is where the system can be able to determine whether the decision is right or wrong. Not only that, it can correct itself when wrong and use the result of its own decision to learn. Deep Learning takes the issue of human supervision out of the equation. This means the computer can learn, make decisions, and perpetually continue to make better decisions from previous decisions. This means overtime, the system can rewrite its own code in a much more efficient way beyond what any single person can think of in a given period. The staggering part of it is that it continues to receive data which makes it learn what humans know without limits (maybe data storage will be the limit) with humans seemingly unable to learn at the same pace.
If you have seen the TV series, Person of Interest, an extreme use of AI is exemplified there. The fact is, almost everything exemplified of the AI in the series is actually possible (in theory mostly for now). Sounds disturbing? Relax, there is so much more to life. Maybe it’s time we discover that man is beyond an intelligent higher animal.
I’ll be here to simplify another tech stuff soon. Cheers!