Alexa, wake me up at 8 in the morning with music! or Hey Google! What is the fastest route to the supermarket?  That's how simple it is to give commands to a smartphone, yes. Suddenly we find ourselves talking to a device and on top of that it responds to us and provides suggestions, in case we are not satisfied. As if by magic, artificial intelligence works wonders in all kinds of places and moments: on your mobile device, in your washing machine, on the GPS, in your car and in countless everyday places.

If you thought that artificial intelligence was foreign to you, then welcome to the 4th industrial revolution.

What artificial intelligence is all about

To delve a little deeper into this topic, we can start by defining artificial intelligence (AI) as the science of making things smarter, whether it is a process or something that a machine develops to provide a result. In other words, they are prediction and recommendation systems that try to emulate human intelligence.

For mathematicians, AI is nothing more than an optimization algorithm that seeks the best way to accomplish a specific task taking into account certain information.

The development of this technology is based on the intelligent use of data availability and, with it, the creation of algorithms that allow us to improve the application of said intelligence. In this sense, what really attracts us and is revolutionary is that access to data has grown exponentially, giving rise to much faster and cheaper processing. This invention turned out to grow by leaps and bounds thanks to cloud computing, a technology that allows remote access to software, file storage and data processing over the Internet.

It is no small fact that during the last decade, artificial intelligence has been a relentless source of business innovation, since its power is developed to the maximum in daily commercial operations, such as the automation of business processes, data analysis, detection of defects in product manufacturing,  basic interactions with customers, among others.

How artificial intelligence works

As a general rule, artificial intelligence applications tend to “train” computers to match images and data that they can recognize. For example, showing a computer thousands of images of roads or flowers allows experts to create applications that recognize these things and, if necessary, act on them. This type of artificial intelligence has been in use for quite some time and is in everyday use. In fact, we can see it reflected in the security methods that Google shows on our screens when we must select the required object to confirm that we are not robots.

From the point of view of intellectual capabilities, the functioning of artificial intelligence contains 4 levels: 

• Comprehension: Through the simulation of cognitive abilities applied to data, artificial intelligence is capable of recognizing images, texts, voices, information and much more.
• Reasoning: Through logic exercises, this technology is capable of connecting multiple collected information, thanks to the intervention of precise and automated algorithms.
• Learning: This is the case of Machine Learning. Here we deal with systems with specific functionalities for the analysis of data input and for its correct return at the output.
• Interaction: has to do with the way in which artificial intelligence interacts with humans. This is where we can find NLP systems -Natural Language Processing-, that is, a technology that allows humans to interact with machines taking advantage of natural language.

AI, a journey to the future

The next crucial step for artificial intelligence comes with the emergence of OpenAI's GPT-3, which uses models to allow a computer to not only recognize images and patterns, but to generate language, text and images on its own. 

The OpenAI company was created in 2015 by PayPal co-founder Elon Musk and Stripe CTO Greg Brockman. This corporation is described as a artificial intelligence research and implementation company. Your mission? Ensure that artificial intelligence benefits all humanity.

When we talk about GPT-3, we are referring to a language model, this means that its objective is to predict what is coming next based on the previous data obtained. To translate it into something everyday, this new technology makes it possible to write two or three sentences of an article and GPT-3 will be in charge of writing the rest... magical, right? It also allows you to generate conversations, the answers to which are based on the context of the previous questions and answers.

Clearly, the aforementioned language model has needed prior training to become the computing monster it is today. In short, the training consisted of “teaching” this code an enormous amount of information available on the Internet. This includes all public books ever written online, the entire Wikipedia encyclopedia, and millions of web pages and scientific documents available on the web. This data processing is nothing more and nothing less than the massive absorption of all human knowledge existing on the Internet.