First, let’s take each word of the term individually and try to explain it. Artificial refers to something made by humans (in contrast to nature made stuff) and usually is a copy of something natural. So, artificial is about humans trying to replicate Mother Nature.
What about the meaning of intelligence? Well, that’s when the fun begins. We call ourselves “Homo Sapiens”, which in Latin means “wise man” and we like to take pride in being so intelligent in comparison to our fellow peers from the animal kingdom. But what does intelligent really mean? Some of us may think of IQ tests, while others may think of surviving. I propose the following definition: being able to learn and apply what was learned (of course, this is a simplified definition, not a thorough one).
Putting these together we may say that artificial intelligence is “something made by humans that is able to learn and apply what was learned”.
This is just a starting point because in real life things are far more complex. The textbook definition of AI is the study of “intelligent agents”: any device that perceives its environment and takes actions that maximise its chance of successfully achieving its goals. Truth be told, there is no quite an agreement about an exact definition even among AI researchers. So, I think the best way to understand the meaning of artificial intelligence is to discuss about its pursuits.
Even though you may have heard about it recently, AI was founded as an academic discipline in 1956 and along the years it has experienced its share of ups and downs (the history of AI will be discussed in a future article).
The long-term of AI is to eventually carry on any task that a human being can do it. It’s more like reverse engineer our brain and then create an artificial brain that functions the same as our human brain. Since we haven’t yet (fully) discovered how our brain works that’s why this goal is a long-term one.
In the meantime, let’s discuss about more approachable pursuits of AI that the reader may have already heard of (there are more than those discussed here).
Machine Learning (ML)
Tagging people and objects in photos, content recommendation, google search, what about them? They are all examples of machine learning application. The next video to watch showed to you on youtube is based on past data of your online behaviour on youtube. Same thing applies to facebook feed algorithm. The more I click on real estate links, the more facebook will show to me sponsored real estate pages. When you search on google let’s say “java” it can show you the first results about coffee or about the programming language depending on your search history. This is all machine learning. It’s about using data to answer questions, finding and extrapolating patterns.
Let’s take a numerical example. We have a couple of values for two number and the task is to find the relationship between them (try to find it alone before reading the answer):
x = 0, 1, 2, 3, 4, 5 y = -1, 1, 3, 5, 7, 9
Using machine learning we can quickly learn that the relationship between the two numbers is the following: 2x - 1 = y. Basically, machine learning helped us finding a pattern in the numbers (data) we have.
Natural Language Processing (NLP)
Siri, Cortana, Alexa, these are prime examples of NLP at play. Natural Language Processing is about machines understanding what you mean when you are saying something, to get the context of the conversation. When you ask Siri to play some rock you are referring to play some rock songs, not to play with a (physical) rock. That’s the challenge of NLP, to make a machine to be able to talk with a human being by understanding what is being discussed and to come up with its own opinions.
Self-driving cars are an example of robotics. A self-driving car takes data from its environment, process that data and makes a decision. Or take for example a smart vacuum cleaner robot. What it really does is to take cues (data) from the environment (your room), process it and decide which way to move. If is continuously bumping into your furniture then that wouldn’t be too intelligent, right? Another example of AI in robotics could be drone delivery. The road from factory to destination is paved with obstacles, so the ability to make the delivery is a matter of understanding the environment and make decisions accordingly.
Coffee mugs inspection on a production line is an example of computer vision. It “looks” at a coffee mug and tries to find any evidence of cracks, if none found then the mug is ready to be packed and shipped to customers. Basically, computer vision deals with making a machine to gain understanding from an image or video. That’s what happens when you use Face ID on your iPhone, the phone recognises you and unlocks itself. This is called facial recognition and is widely used in China where cameras are almost everywhere and can track your behaviour on the streets. It has the ability to catch you jaywalking, information which will be used to lower your social score (which is kind of a reputation score).
AI is comprised of quite a few number of subfields making it a complex and a universal field. To be truly understood you have to see it through its numerous lenses. That’s why is so hard to come up with an encompassing definition.
However, I will end by giving a shot at defining AI in my own view. For me, Artificial Intelligence is the field that deals with replicating every form of human intelligence with the purpose of creating a machine capable of acting and thinking intelligently.
I hope you find this article useful and it will get you intrigued about the AI field. If I manage to do that, then my goal is achieved.
P.S. I’m an indie maker and I’m writing a book on the basics of AI. If you want to support me and you’re interested in AI, then you can pre-order my book at a discount here (you won’t get charged until I finish the book): https://gumroad.com/l/SXpw/sideproject