The artificial intelligence (AI) is a a branch of computer science that focuses on the development of systems and programmes capable of performing tasks that would normally require human intelligence.
Broadly speaking, artificial intelligence systems They can perform tasks that usually require the use of human cognitive faculties, such as processing spoken language or identifying recurring patterns.
To provide the artificial intelligence behind these skills, You need to provide it with a considerable amount of data in which the AI You will need to draw patterns that take an input (input data) and produce an output (result, output data). The aim is to refine decision-making by artificial intelligence systems so that it can provide the best possible answer to the question asked.
In many cases, it is still humans who oversee the AI’s learning process, sorting its responses into correct and incorrect ones. However, There are currently some highly advanced artificial intelligence systems that correct themselves and require far less human intervention.
An example of this is artificial intelligence in video games (TAS), which attempts to complete a video game over and over again until it finds the best way to reach the end.
Types of artificial intelligence
We recognise different ways of applying artificial intelligence and of utilising it to develop AI programmes for various purposes:
General artificial intelligence, or AGI
The artificial general intelligence (AGI), also known as strong artificial intelligence, is a machine capable of solving problems and handling requests it has never encountered before and for which it has never been trained. To date, AGI remains a pipe dream to which AI researchers which we can see in some films featuring futuristic elements, such as *Star Trek*.
As we have said, no form of AI, however intelligent it may be, comes close to the cognitive abilities developed by a human being. AGI does not yet exist because AI cannot devise a satisfactory response based on something it knows nothing about – that is, something for which it has not been trained using vast amounts of data.
Limited artificial intelligence or ANI
The Narrow AI, whether specialised or weak, is trained to respond to requests within a clearly defined context that solves problems within a specific subject area. For example, driving a car, transcribing human speech and curating content on a website. As it is hyper-specialised, Many limited AIs are extremely functional and rigorous, albeit with many limitations.
Some examples of narrow artificial intelligence include:
- Voice assistants such as Siri or Alexa.
- Autonomous vehicles.
- Google Search.
- Conversational chatbots.
- Spam filters.
- Recommendations on film streaming platforms.
Text-based artificial intelligence
The text-generating artificial intelligence systems, such as ChatGPT, are able to produce text independently after being trained on large amounts of linguistic data. These models use the deep learning and neural networks to learn linguistic patterns and structures from existing texts.
When you interact with ChatGPT, you provide it with a context or a question, and the model uses its prior knowledge and training to generate a response that fits the context provided.
Although it is the best known, ChatGPT It is not the only text-generating artificial intelligence. However, it is currently the one that best meets users’ needs. There are other artificial intelligence systems focused on text generation, such as Bing AI, Rytr, Bard AI and Jasper.
Artificial intelligence that creates images
A artificial intelligence A system that creates images is known as an image generator. It uses advanced deep learning techniques, such as Generative Adversarial Networks (GAN) or Variational Autoencoders (VAE), to generate new images from existing data.
This type of artificial intelligence has revolutionised the world of design, as it is capable of generating original images whose artistic quality is comparable to that of other highly acclaimed works of art. The best-known image-based artificial intelligence systems with Midjourney, DALL-e and Stable Diffusion.
Español 
English (UK) 
Français