Artificial Intelligence in Simple Terms

🤖 Artificial Intelligence in Simple Terms

Understand how it works and master it in less time than it takes to read this article.

Generative artificial intelligence (AI, from now on) is changing the way we work, research, test concepts, and can be used today to do practically everything. In this small guide, I will explain in the simplest way possible some key concepts to understand generative AI and how you can get the most out of it.

What is Generative AI?

Generative AI is a field of artificial intelligence that focuses on creating models that can generate new content, based on a set of training data. It's like teaching a robot to paint; first, you show it thousands of paintings and then you ask it to create its own work of art. In this way, all current generative AI models are trained.

But... What is a model?

A model in artificial intelligence is like a human brain, it is composed of a neural network (I know, we keep introducing more concepts, but that's the idea, that you leave with everything you need to understand AIs 🤖) that is capable of understanding some input data and with it generate an output. Depending on the architecture and training data, the input data can be, for example: text, sound, images, videos, and any type of data you can imagine, while the output can also be any type of data.

How does AI know what to generate?

When you use AI, you expect that (for practical examples, let's imagine the case of an image generation model) if you ask for a red apple, it gives you a red apple and not a green pear, but, how do models like Midjourney or Dall-e-3 know what a red apple is? Simple, they don't. We cannot imagine an AI model as an entity that knows something, AI is trained with input and output data at the same time, that is, it is given a text that says "red apple", and it is shown the successful case which is an image of a red apple. In this way, and when you multiply by thousands or millions the amount of training data, the neural network is able to adjust its weights and produce results that are increasingly better according to the quality of the training data.

Wait, are you telling me they have neurons?

Well, good that you ask (although maybe you didn't), but yes, in one way or another a neural network works similarly to how ours do, only instead of chemical signals, they are mathematical operations that pass between them, these mathematical operations depend directly on their weights, and what are these weights? We could describe the weights as numerical values that indicate whether a neuron is more or less related to another according to the input data. In this way, depending on the input data, some neurons will be activated and there will be a determined output.

Let's open a debate, does AI have intelligence?

It may seem redundant, because of artificial intelligence, but AI does not have intelligence in the human sense of the word. The "intelligence" of an AI refers to its ability to perform specific tasks efficiently through data processing and machine learning. It does not possess consciousness or understanding of its own. All its results are determined by the training of the neural network, AI does not think like us, because it does not even do so. AI is, basically, a machine trained to perform specific tasks with precision.

Finally, a little joke produced by my artificial intelligence: Chayito.

Why doesn't AI work well in the kitchen? Because it always tries to find the shortest path to the cake.

Yes, one of its weaknesses is the lack of a sense of humor, although to be honest, humans are not exactly the best humorists for the most part, nor how to blame AI. Oh, and if you want, you can try Chayito at this link PS: It will tell you a joke about cats and technology

If you want to find out more about artificial intelligence, you can follow me on Linkedin or Youtube

My name is Charlytoc, and I hope you have learned something new today.