Artificial intelligence is probably humanity’s most complex and astonishing creation to date. Since 1950s, there has been a lot of progress, but we are still in the early stages of AI, there are still many areas to explore and more are in the making .We still have decades to go before we can transform the theoretical concepts of AI into full-fledged systems.
We will see in this article the 7 artificial intelligence systems divided into 2 categories: functionality and capability.
AI based on functionality
1. Reactive Machines
These are the oldest forms of artificial intelligence systems. They have an extremely limited capacity. These machines do not have memory-based functionality, they do not have the ability to “learn”. They simply react, according to the parameters provided. These machines are only used to respond automatically to a limited set or combination of inputs.
A popular example of a reactive AI machine is IBM’s Deep Blue, created in 1997, which defeated chess grandmaster Garry Kasparov. Deep Blue is able to make predictions about the next moves it and its opponent might play and chooses among the next possible moves.
2. Limited Memory
Machines with limited memory are machines that have the capabilities of purely reactive machines and are also capable of learning to make decisions. It is one of the most stable and frequent types of AI. Most deep learning techniques are memory-constrained and almost every existing application we know of falls into this AI category. This is the case with chatbots, virtual assistants or even self-driving vehicles.
3. Theory of Mind
While the previous two types of AI are still abundant, the next two types of AI are for now either a concept or a work in progress.
The Theory of Mind AI is the next level of AI systems currently undergoing innovation. The realization of the theory of mind level AI requires the development of other branches of AI. It is the AI that would be closest to human behaviour if it ever existed. It would have the ability to perceive the needs, emotions, beliefs and thought processes of the individuals with whom it reacts, but also to express emotions.
Sophia from Hanson Robotics is an example of the theory of mind AI. She can see thanks to the cameras present in Sophia’s eyes, combined with computer algorithms . She can sustain eye contact, recognize individuals, and follow faces.
Self-awareness is the final stage in the development of AI, which currently exists only hypothetically, as an AI that has evolved to be so close to the human brain that it has developed self-awareness.
The creation of this type of AI will always be the ultimate goal of all AI research. This type of AI will be able to understand emotions from those with whom it interacts, but will also be able to have emotions of its own. This is the type of AI seen in the wires where robots take control.
AI based on capability
5. Artificial Narrow Intelligence (ANI)
ANI is one of the most frequently experimented types of artificial intelligence. It is the simplest of the three types of artificial intelligence based on capability In low AI, the focus is on one, or at most two tasks.
Narrow Artificial Intelligence refers to AI systems that use human-like abilities to respond to tasks. This type of artificial intelligence is also known as weak artificial intelligence because it operates with a limited set of constraints. These machines can do no more than what they are programmed to do and therefore have a very limited range of skills. These systems are all reactive AI with limited memory.
The IBM Watson supercomputer is an example of narrow AI. It uses cognitive computing, learning and natural language processing to process information and meet your needs.
There is also google translate, image recognition software, recommendation systems, spam filtering and Google’s page ranking algorithm.
6. Artificial General Intelligence (AGI)
General Artificial Intelligence is the ability of an AI agent to learn, perceive, understand and function completely like a human being. AI systems will be just as efficient as humans, by building independently, multiple skills in different domains, to massively reduce the time needed for training.
The Tianhe-2 supercomputer was developed by China’s National University of Defence Technology. It holds the record for cps (calculations per second) with 33.86 petaflops (quadrillion cps).
7. Artificial Superintelligence (ASI)
ASI systems will be the pinnacle of AI excellence, the most powerful form of intelligence on earth. ASI systems would be better than human intelligence. ASI will be extremely powerful in everything it does, thanks to extreme memory, data processing and analysis, and fast decision-making capabilities. The development of AGI and ASI will lead to the singularity scenario. And while the possibility of having such powerful machines seems attractive, these machines can also threaten our existence, our way of life.