Defining Artificial Intelligence - Ep. 01

This is part 1 of 18, in a series "The basics of AI: A Self-Study Journey"

Artificial Intelligence, or as you’d probably say - AI, is everywhere now. And it’s transforming our world in ways we never imagined.

But what exactly is AI?

Let’s start with breaking down the word: AI = Artificial intelligence.

• "Artificial" is something human-made eg, it’s not found in nature.

• "Intelligence" is learning, understanding, and solving problems.

According to Stanford Prof. Christopher Manning, the term Artificial Intelligence was coined by emeritus Stanford Professor John McCarthy in 1955. McCarthy defined AI as: “the science and engineering of making intelligent machines”.

This is the best definition I’ve heard. Simple and clear. But does it help me, someone without a tech background to understand what AI means? What’s the science? What’s the engineering? What is an “intelligent” machine?

The definition of AI is non-static. 50 years ago, automatic methods for search and planning were considered to be AI. Today, this type of system is so every day that there considered to be basic computer science. Similarly, methods for processing unsure data were before considered AI, today this would fall under statistics and probability theory. 

But let’s take a step back…

How is AI used today?

In elements of AI, they share three examples I’ve summarized here:

Self-driving cars use AI to navigate, warn you about obstacles, and plan routes. Streaming services like Spotify and Netflix personalize content using AI, showing different content to individual users. Ever checked out your mom’s Netflix? Even Face ID on your iPhone is powered by AI - and soon your iPhone will have AI everywhere.

Now, let’s look at why it can be hard to define what Artificial intelligence is:

1. There is no official definition of AI. Since it’s constantly evolving. An old joke says AI is "cool things that computers can’t do.” As you probably figured out by the joke - this means the definition of AI will always change since computers constantly improve.

2. I don't know about you but I love good sci-fi. If you do too - then you’ve seen one or two human-like robots on the screen. Fiction shapes how our view of AI. It even pushes some scientists to try things they've seen in movies as kids.

3. What is difficult for us, might not be for a computer. And what is easy for us, might be incredibly hard for a computer. For example, AI’s intelligence has often been measured as whether it can beat a Chessmaster in chess - but chess is pretty straightforward for a computer.

AI as a Discipline

AI should be considered a broad scientific field/discipline, rather than something that can be quantified as individual entities (eg "one AI" or "two AIs"). Referring to "an AI" is inaccurate; instead, we should talk about "AI methods" or "AI systems."

In other words: AI is a scientific discipline comprising various concepts, problems, and methods.

Final thoughts

In summary, AI is a broad scientific discipline that encompasses various methods and systems. Referring to "an AI" is incorrect; instead, we should talk about 'AI methods' or 'AI systems.' It is important to understand the basic concepts, it helps us understand how AI can and will affect our everyday life and our future. I think it's important to break down the meaning of AI, even the word itself.