True AI

Artificial Intelligence by definition is getting redefined as it is getting evolved. In-fact, AI has created its own history in this evolution. A unique history that started for the very first time the term AI was coined in 1956 by John McCarthy till today. And it does not stop here – Everyone of us are carrying AI in our road map in some form factor towards our future.

MIT Professor Max Tegmark in his recent book – “Life 3.0: Being Human in the Age of Artificial Intelligence” captures the evolution from a different dimension – He lists the evolution as below –

Life 1.0 – Biological
Life 2.0 – Cultural
Life 3.0 – Technological

Technology is empowering every discovery and innovation. In the same way high computing power and intelligence are being built to make AI truly possible.

In this article, let us see the broad classification of AI itself and how different use cases fall under these buckets.

ANI (Artificial Natural Intelligence)
AI at this level specializes and focuses on one use case or category. They know how to react for a specific situation. They understand context within their known boundaries. The context are not cross-functional.

For example take a self driving car – They are specialized only in one category – Driving with caution – They are equipped with only enough information to achieve a specific goal.

I see AI at this level are different entities that can do parts of human task with intelligence.

AGI (Artificial General Intelligence)
When machines can think, act and respond like humans it is called AGI. This is a strong AI or Human-Level AI. Response to stimuli is that of humans. They compete and work with humans as one among them.

Let us take driving as an example here. Rather than car being driver-less car – If there can be a humanoid robot which can drive today’s car and also ride a bike and similarly do varieties of task then AGI is achieved. Here humanoid is equipped with the same intelligence as humans and is trained to handle multiple categories.

AI at this level are self-aware, self-trained like humans.

ASI (Artificial Supernatural Intelligence)
The capability of machines surpass the intelligence and capability of humans. This is a very futuristic approach. Easily you might compare hyper automation with this and get into a conclusion that machines are already surpassing human capabilities. This is a wrong conclusion. Hyper automation is about doing the same mundane task in much better way eliminating the precious human talent in doing it. ASI is about speed with intelligence. They are often considered as super-humans or mythological demigods.

To be honest – Speed and accuracy in anything humans can do is one of the factor that can be considered as super human skill or leap towards next level human.

They have the capability to foresee the future and help humans to equip or train towards it.

Where are we on all these above levels of intelligence?

Interestingly the error rate of an trained intelligent machine focused on a particular task has come below the trained humans.

All the above types try to break the misconception that Intelligence is not single dimensional. It is multi-dimensional and multi-variant. And what we have today is just a start!