It's been 22 days on steemit and I am looking forward to doing my 1 month journal soon on what I've learnt blogging on this revolutionary platform. However, for this post, I will like to keep my promise and start a series of articles on what I call AI school. In these articles, I'll be discussing several areas of AI but at a level at which almost anyone can understand. One of the sins I think the AI community has committed (and one can argue still committing) is always treating their work as too complicated for the layman to understand. The challenge with is of course is that once the so-called layman comes to understand and appreciate the specific AI problem or solution, then the temptation (for the layman) is to conclude that such a problem or solution was not AI in the first place. This is of course the classic curse syndrome and I talked about this in my previous article on Autonomous (AI) driving no longer fiction. I believe this hurts the AI community and I've decided to choose a different path - my articles will be definitely AI and they will be easy to understand. So let's start with the first of the series, befittingly enough, what exactly is AI?

Source: Image by OpenClipart-Vectors from Pixabay - Time to go back to (AI) school.

What is AI?

Artificial Intelligence (AI) has several definitions, and without boring you with theoretical or academic definitions, I'll just define AI as any machine based system that can think or act in the world without being explicitly programmed. There is a lot to unpack here but I'll concentrate on the words in italics.

Any machine based system:

It is not surprising that a human being or animal cannot be considered AI. Only machine based or derived systems can be considered AI. This means that your TV can be considered AI but your dog cannot (not matter how intuitive it is).

Can think:

Examples of AI systems that can think include Apple's Siri, Amazon's Alexa, IBM's Watson e.t.c. Notice that these systems are largely computer programs or software and are not physical systems (Siri sits on an iPhone but Siri is not the iPhone). Herein lies one of the biggest misconceptions about AI. AI doesn't have to be a visible physical robotic system but can be an invisible face detection software monitoring your every move.

Can act in the world:

Of course, AI can also be a robotic system that can perform certain tasks. For example, in manufacturing, robotic arms have increasingly been used to perform routine, repetitive tasks as simple as such as "picking* or as complex as assembly. Also, their actions must be in the world too. For example when you ask Alexa a question, it thinks when it is interpreting your question and searching for an answer in its vast data base and acts in the world when it talks back to you with the response.

Without being explicitly programmed:

This is by far the most confusing part of any AI definition and its quite understandable. For example, almost every one knows Siri and Alexa were both programmed (and probably still being programmed as we speak) by armies of software engineers but the key point to remember is that not all its thinking or acting is being programmed. This means that there is a significant learning component, on its own, that makes Siri and Alexa qualify as AI. Such learning is generally - and appropriately so - called machine learning (ML), a sub-part of AI that will be discussed in subsequent school sessions.

So that is AI in a nutshell and congratulations in completing your first class in our AI school! Subsequent school sessions will deal with several aspects of this fascinating technology that is changing our lives at such a rapid pace.

Next School Session

The next school session will be on the different types of AI - Narrow AI and General AI - which promises to be quite interesting. Finally, if there are any aspects of AI, you might want me to touch on, please leave a note in the comment section and I'll do my best to address each of them in subsequent school sessions. Subsequent sessions will apply a demand first model, meaning I'll address topics that are in demand from you, my readers, unless of course I think there is a pre-topic that might help to better understand the current in-demand topic.

Further Reading & References

For those of you that are academically inclined, Russell and Norvig's AI: A modern approach is a dense but very good read.

For those that are more visual learners, deeplearning.ai Coursera AI for everyone is a fantastic introduction to AI for almost anyone.