Journey Through The Various Stages Of AI Evolution [Self-Aware AI; In-view]

Introduction

The term "Artificial Intelligence"; AI, is not entirely new in this era, except for the uninitiated. And just as you may know, there have been amazing advancements in this field which has been indicated in the designing of AIs that have thrown the world in awe. In this survey, we will be looking at the "past, present, and future", and of course, the ultimate destination of Artificial Intelligence; which is the "Artificial Consciousness". I would also give you some real-life examples of each stage of evolution of AI.

[Image Source: Maxpixel. CC0 Licensed]

Before I proceed, may I remind us that one of the purposes of AI is the creation of Machine Learning system that can mimic; or rather simulate; the human domain of learning (cognitive domain, psychomotor domain, and efferent domain) in an unnatural way.

But as it were, we have only seen the simulation of the cognitive domain, while the other domains have been left in the grey area. But hey! Machine Learning (ML) has been breaking bounds lately, and we will be having other domains fully represented in AIs. But let's look at where it all started; the fundamental types of AI; then we will be extrapolating it down to the future.

1. Reactive Machines

This; in my opinion; is called "AI at its ground state". This type of AI is one of the earliest attempts to incorporate human-level intelligence in machines; but only in its crude state. This is because, this type of AI does not relate past experiences with the real-life scenario as they present themselves.

Okay! In a layman's term, this is a bunch of algorithms that is designed to perceive the immediate surrounding/environment and perform activities in relation to what it perceives; hence the name "reactive AI". Note: This AI cannot form new memories, cannot think, and cannot draw inferences from a pre-defined database of past experiences.

Example of the reactive AI is the Google AlphaGo, which was seen to beat the human champion at the game of GO. But amazingly enough, the AI was not even aware that the game of GO is actually a board game, according to the report here.

[Image of AlphaGo V.S Sedol from flickr. Author: Buster Benson. CC BY-SA 2.0 Licensed]

Now you see what I'm talking about, this type of AI is only stereotypically programmed to perform a pre-defined task. Like the AlphaGo cannot write a love poem for you. Okay, this would lead me to the next evolution of AI.

Limited Memory AI

Just as the name suggests, the AI is incorporated with some basic and fundamental memory database. And it is programmed to relate to the past; and/or database; and use it to solve a particular real-life problem or to respond to a real-life situation.

Examples of this type of AI are the autonomous-driving vehicle, chatbots. One very unique thing about this AI; which is not present in the reactive machines; is the fact that it can collate fragments of past experiences/data and use them to make inferences in relation to solving a real-life problem. For example, the autonomous-driving vehicle would know when to stop at the instance of the red traffic light and would know when to move.

[Image Source: Pixabay. CC0 Licensed]

But it is worthy to note that this AI is also devoid of both common sense and consciousness; and can only simulate some parts of the cognitive abilities of the human brain. This also forms the borderline between the weak AI (narrow AI) and the strong AI (Artificial General Intelligence). Okay let's look at another evolution of AI.

Mind-Theory AI

This is actually a major breakthrough in the field of Artificial Intelligence. This type of AI has the ability to perceive thoughts, understand them, and use them to relate with humans. Here's one awesome fact about this: They can detect humans’ emotions and behavioural pattern.

This has gone beyond just the cognitive domain into the efferent domain of learning, which includes emotions, thoughts, and some are even designed with some levels of feelings. And yes! They can interact intelligently with humans. I'm sure you must have seen the "Star wars" Sci-fi; the R2-D2 is a perfect example of this. Hey! Don't be quick to say it's fiction; sci-fi is now turning to sci-fact.

Majority of the AIs represented in this type are called the Artifical General Intelligence (Strong AI or "Full AI"; according to how Ray Kurzweil refers to them), which is an enhancement; or rather, a step ahead from the normal AI. This type has the ability of cognitive computation just like the human brain.

No doubt, this has raised some reservations by people because of the possibility of it evolving with full consciousness and becoming self-aware. They fear that they could lead a resistance against humans one day. This was exactly what the recently demised professor of theoretical physics spoke about:

AI would take over and end the human race. - Stephen Hawking. [Paraphrased by me] Source.

The AI he talked about was the self-conscious AI, and in this contest, it is an evolution from the Artificial General Intelligence. Now, this leads to the ultimate destination of AI; okay let's find out where the destination is.

Self-Aware AI

Yes! This is the final destination of the technology of AI. At the moment, this only exists majorly theoretically or hypothetically (but you should also be aware of the fact that; something being theorized doesn't equate it as fiction). This type of AI can be said to have evolved with full consciousness which is comparable to; if not superseding; the consciousness of the human mind.

[Image Source: Pixabay. CC0 Licensed]

The qualities of this AI include: Sentience, self-awareness, emotions, thought, super-intelligence, feelings, has the abilities of drawing inferences from real-life scenarios, and most importantly "common sense". This was the type of AI that was implied when the futurist; Ian Pearson made this witty statement:

AI could be billion times smarter than humans. - Ian Pearson [Source]

Example of AI in this category is @samminator (I'm kidding). In reality, there hasn't arisen any fully self-aware AI yet, but I can draw you to the sci-fi; Ex Machina; the "Ava" in that movie is a perfect depiction of this. But with further advances, we could see conscious machines arising sooner than we imagine. Remember, tech is dynamically evolving.

Conclusion

The technology of AI is rapidly evolving. We have seen the various stages of this evolution and what AI will potentially evolved into. But one thing I would love to state here is that there is no limit to how advanced AI can be, the only limitation resides in the human mind. As long as the mind can conceive it, the actualization would just be a question of time

Thanks for reading

References for further reading:

• Computational-Intelligence.pdf
• Technology-review/ComputerConsciousness
• Betanews/Types-of-AI
• HowStuffWorks/Robotics-AI
• LiveScience/Types-of-AI

All Images are CC Licensed and are linked to their sources

gif by @foundation