What is AI, and what is not?  

Many people don’t know what AI means. But, unfortunately, you, too, could be misinterpreting what it means.

AI is so often used that it’s hard to go through a week without hearing or seeing the word ‘AI.’ Even I have joined the trend and been tossing the word without genuinely understanding. I used the word carelessly in sentences seeking investors, even when I was featured in the media. Only to find out that I had it all wrong.

Many of us are using it without really understanding what it means. And if you are like me, I learn vocabulary by its use in the sentences. I don’t quite understand the origin or true meaning until someone points it out.

It’s like the term “money shot.” I used the word loosely after great success with something I’m doing well. Until someone politely pulled me aside to stop referring to pornography.

Understanding the true meaning of a word can change how others perceive us. So, especially when we don’t use the term correctly, let’s get to the root of what artificial intelligence, or AI, means.

To think it has to do with automation is wrong.

Anything that has to do with automation gets lumped as AI. It’s what the media and the general public say to communicate AI. This is incorrect.

When you are about to talk to a developer or someone to help gain support for your AI-led product, like an investor, you need to take it to the next level and communicate better so that they understand what you are ultimately trying to achieve.

Artificial intelligence, or AI, is an umbrella term comprising various techniques and approaches to building a digital application that can execute a specific task. The way you use a computer to complete a task, each has its approach. So that is why you must understand what problem you are trying to solve and how an AI system can help with solving this problem. 

What is AI really?

People often say that AI is a computer system or application that can mimic human behaviour and thinking. When you see the word intelligence in the term artificial intelligence, you immediately start to think about superior intelligence systems that mimic the human brain. Computer systems or applications that can reason and intelligently respond. That’s not precisely what AI means; the intelligence term is misleading.

AI can only predict.

AI systems can predict one or millions of things combined to create intelligence similar to humans or the perception of a task being automated. However, AI is not automation by itself.

Here’s an analogy that you can use to help better understand what AI is. AI is like a cell. A single cell has a function to perform a specific task or, in a technological sense, predict. It can predict what is happening now or in the future. It makes its prediction on the information or data that are present.

A cell can work with other cells, and they can make up a tissue together. A group of tissues gather to make up an organ, which essentially operates to perform a function in the human body. Organs are what we see and understand. For example, my heart beats to pump blood throughout my body. Each cell in my heart has a specific role.

In an artificial intelligence system, prediction occurs. For example, when using a facial recognition app, the AI predicts the likelihood that a particular face is a match based on a collection of previous data points. I can add another layer by taking the outcome of a prediction or multiple predictions together. The new layer can further the function of the application to do a particular task based on the result of various predictions. This creates the perception of automation.

Decision-making systems, automation, and statistics are not AI

If AI can only predict, then decision-making systems, automation and statistics are not necessarily synonymous with AI.

In a decision-making system, if you have a bunch of information that requires sorting, it isn’t an AI system. It’s not predicting anything. If you have a system that needs to rank data according to specific criteria, then it is not an AI system. Again, it’s not predicting anything. If you have an application that requires you to choose among various information based on a pre-defined set of rules or logic, you also do not have an AI system. No prediction of the current or future event has occurred.

When there are no rules or logic available, or one way to approach the problem,  at the start, but you need the computer to learn what the rules or logic could be to predict an outcome, then AI can apply.

Machine learning and deep learning are sub-division under AI.

Under the umbrella term AI, we have machine learning and deep learning. Each has its purpose. In machine learning, we have a system that is learning from the data that we provide to the system to study. The data it learns becomes a template for other information given to the system to help predict the possible solution or outcome.  The template gets improved with human intervention. When new data is added, correct and improve the template to become more accurate or predict better in the future.

Machine learning is used in many recommendation applications like Spotify, showing what music you might like based on the music you’ve listened to. It can also be used for predicting the type of videos you would like to watch based on the videos you’ve seen. Whether it is correct or not, it is based on whether you click to watch or listen.

Deep learning is a subdivision of machine learning that starts to look more like human capabilities. So it is when a system is learning to make decisions independently. How we interpret and understand information is the same way that the deep learning system can be designed to operate.

Your choice of how you decide what you use for training can be the difference between machine learning and deep learning. This is because the data in deep learning is not organized or labelled according to what you want the computer to learn. Instead, deep learning allows the machine to infer what it thinks is important and reason independently.

Which is why AI can be dangerous. The choice of data presented to the machine can make a huge difference in how it decides.

Use the term the right way.

You now have the basis for understanding what AI is. You can now focus on explaining what product you want to build. Be clear if and when you’re using AI and when you’re not. Finally, you can demonstrate your understanding and respect from others in the field to join your mission to execute your vision.