Artificial intelligence is in the making for more than half a century, with small incremental steps from the early 1940s. Like any invention or technology, the growth in research and innovation in the field of AI is exponential as we progress. To say that AI is still in its nascent stage is just astounding, thinking about what we have already achieved with AI.
What is Artificial Intelligence, although it is very obvious from those two words, well actually just the second one; intelligence, according to Wikipedia is the “capacity for logic, understanding, self-awareness, learning, emotional knowledge, reasoning, planning, creativity, critical thinking and problem-solving”. Wow! It is just overwhelming when you think of how big each of those topics is in themselves. Intelligence is what we humans create is referred to as Artificial Intelligence. Almost all aspects of intelligence have been explored and are at various degrees of evolution. This splits up AI into a few types that help us to explore them individually. Let’s take a look at these types, shall we?
Types of AI
There are 2 types of Artificial Intelligence classification systems as yet. One which is based on the ability of AI to mimic human intelligence and self-awareness. The other based purely on the ability to learn and perform tasks. Let’s look at the first kind of classification and the types it enumerates.
The first type of classification which is based on how well the machine is able to mimic human intelligence and self-awareness categorizes AI solutions into
Having no memory-based functionality, these machines were built to respond to various kinds of stimuli, much like the human body. The earlier forms of AI were reactive kind, with limited capabilities. With no memory, there is no ability to learn and recall past experiences and hence no learning capability. These machines can be used for automatically responding to a limited set of stimuli. A simple example is the robotic vacuum cleaner now available for homes. As a complex example, you have IBM’s Deep Blue that had famously beaten chess Grandmaster Garry Kasparov.
Limited Memory AI machines are Reactive Machines with learning capability added. Experiences are stored in memory, also called historical data, and algorithms use this historical data to make decisions, basically implementing what was learnt. Deep Learning systems are Limited Memory AI systems that are trained by large volumes of data, helping to form a reference model to be able to solve similar problems in the future. They are limited in a way that; they will learn from only the data they are designed for. Image recognition is a widely used example of Limited Memory AI. Other examples of limited memory AI include chatbots, virtual assistance or self-driving vehicles.
Theory of Mind machines are systems that exhibit a certain level of emotional intelligence, being able to understand emotions, desires, beliefs, aspirations and maybe intelligence of another machine or human. If you have to envision a Theory of Mind AI machine, remember the robot TARS in the Hollywood flick Interstellar. There are many other such examples in Sci-Fi flicks.
Self-Aware AI machines, although not a complete reality yet, will be systems that understand the concept of emotions, consciousness, beliefs and desires of their own. These kinds of machines usually trigger a fierce debate on how safe these machines will be for humans, especially when they see the human race as a competition and a threat to their existence. Self-Aware machines will be much like humans but with the machine-like capability to carry out complex work.
The second type of classification, based on the ability of an AI machine to understand and perform tasks, classifies AI systems into 3 further types.
- Artificial Narrow Intelligence
This type of AI machine evolves to do only a certain kind of task autonomously, mimicking human capability. These machines have a narrow range of capabilities, hence the name. All Reactive and Limited Memory AI systems can be considered to be part of Artificial Narrow Intelligence.
- Artificial General Intelligence
Artificial General Intelligence machines will have capabilities and competencies much like humans. Yet under research, these machines will have the ability to understand or learn intellectual tasks, just like human beings can. AI prevalent today is said to be decades away from Artificial General Intelligence. Estimates of time needed to build a truly flexible Artificial General Intelligence machine is anywhere between 10 to over 100 years.
- Artificial Superintelligence
Artificial Superintelligence is any intellect that surpasses the cognitive capabilities of humans in every domain of interest. In addition to matching the complex, composite intelligence of humans, Artificial Superintelligence enabled machines will be way more efficient and better at everything they do, with memory and processing power surpassing that of humans.
How long Artificial Intelligence will take to fully evolve and how will it impact humanity is a matter of research that might go on for a few years or even decades to come. But the fact is, businesses are reaping the benefits of artificial intelligence already, and there is an opportunity for anybody with the requisite skills in Artificial Intelligence and Machine Learning to take the lead over everybody else. If you are looking to scale up your skills to a level where you can apply AI and ML powered applications to make processes within your business much more efficient and robust while you improve on your career prospects, take a look at the certification on Machine Learning and Artificial Intelligence course from Jigsaw academy today.