3 min read

What is AI?

Featured Image

Artificial Intelligence (AI): What Exactly Is It?

What is artificial intelligence? How is it used? And why is everyone suddenly talking about it? 

Over the last 7-8 years, AI has taken a firm hold in the public consciousness — we hear about its use everywhere, from self-driving cars and smart home devices, to medical analysis and customer support.

But what is AI, and how does it work across so many different applications?

In reality, artificial intelligence, or AI, doesn’t refer to one specific thing at all, but encompasses a spectrum of technologies. Though AI has been around for decades, it’s only in the most recent handful of years that it has become commercially viable. Even then, we’re just at the cutting edge. It's similar to the story of the steam engine. When James Watt invented it, the steam engine itself was kind of a bust. It wasn't until years after Watt's death, when a new kind of boiler was invented, that the steam engine took off. 

The same is true with AI. While the methodology was invented decades ago and perfected over time, the real boom in AI happened as exponential increases in computing power allowed us to process massive volumes of data quickly.


At its core, AI mimics the decision-making ability of a human – that is, the ability to predict outcomes based on input from the environment. The real value add of AI is that it makes complex predictions efficient and inexpensive

AI replaces skills — not people. It automates time-consuming tasks that are more efficiently tackled by AI technology, which in turn allows humans to dedicate their time effectively to the types of tasks and skills that only humans can do. 

So what are the different types of AI?



In a traditional program, data and rules are fed through a machine to generate an output, or answer.

In machine learning (a subset of AI), a computer is given larger volumes of data (many thousands or sometimes millions of examples), as well as the required outputs, and the computer itself figures out what specific rules lead to those outputs.

For a real-world application of machine learning, we can look at how credit card companies manage credit card fraud.

In this situation, the data given to the computer would be all of the purchasing history on the body of credit cards issued by a company. The output or ‘answers’ would also be given to the computer, and would include the option of a particular transaction being fraud or not-fraud. Finally, the computer would analyze millions of transactions worth of data to predict the rules, such as: where purchases are typically made, the cost of a typical purchase, and the frequency of purchases, to be able to pinpoint deviations in the rules and mark them as fraud.

If you live in California, and all of a sudden a purchase for $5280 made in Croatia shows up on your card without a related purchase (such as a plane ticket), the program will recognize this as a deviation from the rules.


Let’s look at the classic example of sorting dog and cat photos. As humans, we can easily determine a few features that differentiate cats and dogs: overall size, ear shape, and snout shape/length. But this is actually an incredible feat!

In deep learning, a machine is given many thousands or millions of examples of cats and dogs, and the machine itself picks out the features (or rules) without human intervention, using multi-layered ‘neural networks’ that mimic the neural networks present in the human brain. In the ‘training process,’ the machine’s results are compared with the human’s results in order to validate the machine’s attempt at generating rules.

A neural network, in short, is a series of algorithms that aims to understand underlying relationships and patterns in data, and make predictions.


Today, deep learning is used to create efficient and accurate customer support, revolutionize medical care (including detecting cancer cells, and analyzing MRI images), direct self-driving cars, and in Riiid Labs' particular field of speciality, power the future of education in both schools and the corporate world.

The catch? This technology requires a lot more data, and processing such massive volumes of data requires an enormous amount of computing power — a lot more than the average computer can handle.

Implementing AI solutions is a complex undertaking for many organizations across dozens of industries. Without expert consulting and business analysis, the challenges of data shaping, collection, labelling, and storage, lead to limited revenue and profit, and a failure to drive business needs. Especially if you haven’t done it before.

Thankfully, we have.

In fact, it’s what we do every single day.