Categorizing artificial intelligence
Artificial intelligence can be divided into several segments, but two main categories can be identified.
The classification of artificial intelligence systems is divided into two categories: weak AI and strong AI.
Weak AI, also called narrow AI, is an AI system developed and trained specifically for a single purpose.
Examples of this type of AI are virtual personal assistants such as Apple's Siri.
Artificial general intelligence, also known as Strong AI, is an AI system that has the ability to provide human-level cognitive capabilities, which means that if presented with an unspecified challenge, it is capable of finding a solution. .
Developed by mathematician Alan Turing in 1950, the Turing test is a method that attempts to quantify a computer's ability to display human-like thought processes, although its effectiveness is disputed.
Arend Hintze, assistant professor of integrative biology and computer science and engineering at Michigan State University, offers a comprehensive categorization of AI based on four types: from AI systems existing today to sentient systems that do not yet exist.
Specifically, the categorization includes:
The reactive machines
Like IBM's famous Deep Blue chess program.
Deep Blue was famous for his ability to defeat world-famous chess grandmaster Garry Kasparov in the 1990s, possessing the ability to both discern the pieces on the board and formulate predictive responses.
However, Deep Blue was unable to rely on past experiences to influence its future decisions.
You evaluate the available options, both yours and your opponent's, and make the most advantageous decision.
Deep Blue and Google's AlphaGO were built for specific purposes and cannot be adapted to different contexts.
These machines are a simple concept used to simulate human behavior by initiating reactions to stimuli.
They are not designed to learn, but to give an answer.
Another type is limited memory.
These systems are capable of drawing on previous experiences to make decisions in the present.
Many aspects of the decision-making process in autonomous vehicles have been conceived in this way.
The data collected from observations is used to anticipate and adjust to possible future scenarios, such as lane changing, in a timely manner. This information is not kept for a long time.
These AI systems are reactive machines with a certain memory capacity, allowing them to learn data and make small decisions.
Examples of this type of AI are facial recognition systems and chatbots.
Theory of mind and self-awareness
Theory of mind, a psychosocial concept, is the recognition that other individuals possess their own set of beliefs, motives, and goals that can shape their behavior.
To date, this type of artificial intelligence has not been developed.
It is, therefore, a pioneering AI model that aims to improve the understanding of human behavior, such as emotions, needs and thought processes.
Their research and possible applications are very promising and have a bright future ahead.
Finally, self-awareness.
In this area, AI systems possess a form of self-awareness, demonstrating a level of consciousness.
The ability to recognize and consider one's current state allows the ability to understand the emotions and states of those around them.
Currently, this type of AI has not been fully developed.
Although this is just an idea and no AI of this type has yet been created, self-awareness implies that machines, apart from imitating human behavior, could be able to think and act autonomously, demonstrating a sense of self-awareness.