What is Artificial Intelligence ?

By Emmy Sevón, on 14 September, 2020

A report released by the Finnish Ministry of Economic Affairs and Employment states that 15% of today’s jobs will be replaced by Artificial Intelligence (AI). At giosg, we actually think the percentage will be a whole lot higher. However, new jobs will emerge and with AI taking over your everyday tasks, you'll have more time to innovate or do whatever it is that you like to do. The term Artificial intelligence frightens some and fascinates others, this is no surprise since we naturally fear the unknown. 

When I first started my traineeship at giosg, I had no idea what that term actually meant, so I googled “What is artificial intelligence?”. Turns out I wasn’t the only one looking for an answer since hundreds of thousands of people google the very same question on a monthly basis. Now that I’m surrounded by people a lot smarter than me in the matter, we decided to team up and write an 'AI for dummies' if you like, to make the term more approachable.

Firstly, AI is a big umbrella term which includes many different techniques and subsets that people easily mix together

Terms like machine learning and deep learning are subsets to artificial intelligence.

Most of us interact with AI on a daily basis without even realising. The spam filter in your inbox, Snapchat face filters, Google Maps and online shops are just a few examples. Now if a dog filter on Snapchat doesn’t make AI seem less frightening, I don’t know what will.

When googling the term, you find several discussions about what actually can be defined as AI. Well, no


one really has an exact definition and the term is still being debated by academics all over the world.

However, AI does not need to be comparable to human intelligence

This seems to be a general misunderstanding. AI that is comparable, or could potentially outsmart human intelligence is called general AI. This type of artificial intelligence does not exist in the world today, as not even the most advanced robots can be compared to the least intellectual human being. There are too many components in natural intelligence which can’t be traced yet. So no panic, our AI team reassures us that the risk of robots taking over is not a fear of the near future.

Instead, let's concentrate on the type of AI that we see around us online on a daily basis; narrowed or weak AI, This includes what playlists Spotify recommends you, what ads you see on Facebook, asking Siri for valuable advice, all those things we don’t even pay attention to anymore. This is the type of AI that can be taught to perform specific tasks, but does not have its own intelligence.

Even though we don’t have an exact definition of AI, we could say it’s an artificial system that solves tasks requiring intelligence by perceiving its environment. Without a human telling it exactly what to do.

Sounds cool right? Now let's go back and have a look at those confusing terms;

Machine learning is the type of AI that is used in most of the examples we face in our everyday lives. It’s based on a software learning through experience and using this experience to define what will happen next. The more frequent of a visitor you are, the more knowledge it will have on you.

Deep learning, on the other hand, has a more complex structure and is based on attempts to model the brain functionality. The concept is built around the ideas of a 'neural network' , and although it is still a simplified version of real brain functionalities, this approach has been found to be very efficient for multiple machine-learning applications.

Some compare AI to the invention of electricity and we’d like to agree

AI solves various problems and makes many processes more efficient and easy. It assists the marketing team with promoting ads and the sales team with finding prospects. At the same time it’s replacing humans in many industries.

Looking back, innovations should not be held back but let flowing freely. If they wouldn’t we would certainly not be talking about AI or Live Chats today.

"The Americans have need of the telephone, but we do not. We have plenty of messenger boys."
Sir William Preece, General Post Office, 1878.

Little did Sir William Preece know that a messenger boy could be replaced by an AI powered live chat.

How we use AI

At giosg we track visitor behavior, which includes page visits, time on site and returning visits. Based on this behavior we train a machine learning system to build a predictive model. We continuously use A/B testing to test different marketing actions to see how that affects the conversion rate. giosg can help companies whose core business isn’t AI to make use of it in their online business. 

Topics: AI