fbpx
Artificial Intelligence: 4 Levels of Explanation Part 1 (Episode 3 | CVFE)

Artificial Intelligence: 4 Levels of Explanation Part 1 (Episode 3 | CVFE)

You can Watch Video Version of this post below

In the previous post on computer vision, we gave the simplest possible introduction to computer vision and its domains. But computer vision itself is a part of a larger domain known as Artificial Intelligence. Understanding this domain is crucial to be able to connect the dots between different fields in artificial Intelligence. So today we’re publishing a 4 part tutorial/video series on AI as part of our CVFE course, we will focus on giving you a thorough understanding of the artificial intelligence field with these 4 tutorials. Each tutorial is distributed into different levels, on each level we’ll cover some explanation of AI. And on each subsequent level, the explanations will get more technical. 

Artificial intelligence 4 levels

So this tutorial (Level 1) will give you a  high-level introduction to AI but the following posts will go deeper, exploring many of the technical aspects of Artificial Intelligence.

With that in mind, let’s just get started.

Introduction:

Most people develop their notion of Artificial intelligence from watching humanoid robots in media performing crazy stunts, or self-driving Teslas on the highway, or the Chinese smart surveillance systems you often hear about in the news. Perhaps you might have come across Smart stores like Amazon Go and wondered what sorcery is this?

Sci-fi books, movies, and TV series have also built our perception of Artificial Intelligence for the longest time. Movies like Terminator and Matrix introduced us to highly advanced conscious artificial intelligence systems while shows like Black-mirror painted a picture of a dystopian future where Artificial Intelligence dictates different aspects of our life. 

Artificial intelligence Black mirror

All of this makes you wonder what the future holds for us when it comes to AI. 

While many of the incredible AI systems from Sci-fi have been implemented in some shape or form, there is still a long way to go before we expect true consciousness from an Artificial system. What people need to understand is that artificial intelligence systems today can be categorized into one of the following:

  • ANI(Artificial Narrow Intelligence)
  • AGI(Artificial General Intelligence)
  • ASI(Artificial Super Intelligence) 

For now, let’s focus on AGI and ANI, often also referred to as Strong AI and Weak AI.

AGI VS ANI:

Artificial General Intelligence, as you can tell by the name, AGI possesses intelligence comparable to humans. They can perform tasks that require the level of cognitive abilities a human mind possesses. This is the type of AI that you usually see being depicted in movies in the form of characters such as Terminator, who is able to drive vehicles, shoot weapons and use all of his senses to achieve a certain goal. 

Artificial intelligence AGI vs AGI

But you probably haven’t come across such a killer bot yet. This is because, in reality, the progress we have made in Artificial General Intelligence is limited. What you usually hear about in the news media are actually great examples of ANI Systems, Artificial Narrow Intelligence. 

ANI (Artificial Narrow Intelligence)

In comparison to AGI, an ANI system focuses on carrying out single specialized tasks like recognizing faces, monitoring for a restricted activity, or detecting traffic violations. Though the ANI system is limited to performing a single task, it does it really well. 

Artificial intelligence ANI

Most of the progress in artificial intelligence has been focused on building ANI systems which have also been well reflected in the media. 

Artificial intelligence headlines

This rapid progress in ANI is because we have some clearly defined blueprints for developing such systems and the promising results that it yields. The healthcare industry, for example, has made significant advances in developing ANI systems capable of performing better diagnostics than human experts, which is also a lot faster and cheaper. This has the potential to impact and save thousands of lives. 

Artificial intelligence in medicine

As progress in ANI continues rapidly, we will witness more and more such systems deployed in different industries. Hence making an extraordinary impact.

Now you may wonder about virtual assistants like Apple’s Siri or Google’s Assistant on your smartphone, or a self-driving Tesla.  

Artificial intelligence virtual assistants

They sure seem to get a lot of stuff done at a time, so are they examples of AGI? How Can ANI take care of so many tasks at a time?  Take the example of self-driving cars, not only do they steer the vehicle or control speed but they also have to watch out for pedestrians and other vehicles, or process all traffic signs signals on the road, all while maintaining an optimal course to the destination.

Artificial intelligence self driving

It sure seems a bit overwhelming for an ANI system to do all of this and you will be correct to think it’s not possible for ANI. But is it AGI then? Well not exactly, these sort of complex systems are made simply by combining smaller ANI systems that each handle a single task. Though this may give an illusion of AGI, it is simply a number of ANI systems working together.

Artificial intelligence ani detection

Using multiple ANI systems together to do a complex task is hard and perhaps this is one of the reasons we still haven’t witnessed the big claims made by autonomous industry come to reality.

The thing that most people don’t seem to understand is, despite the advances we have made in Artificial Intelligence it’s still really hard for us to mimic human-level intelligence. Even if autonomous cars reach mass adoption, they will be in a restricted environment. I personally don’t expect to see a level 5 autonomous car before a decade. A level 5 autonomous car means a car that can autonomously drive anywhere on the planet without any human intervention. It’s a really difficult task.

Artificial Super Intelligence (ASI):

Another less discussed category in AI is called ASI, which stands for Artificial Super Intelligence. This is just like an AGI that can do all tasks a human can, but additionally, it also possesses intellectual abilities superior to humans. So theoretically an ASI system would outperform a human at any given task. 

Two examples of an ASI system that you may have been familiar with are Skynet from terminator and Vision from Avengers.

Artificial intelligence ASI

ASI is also said to be capable of self-awareness meaning it can develop a consciousness. 

and this is the same AI that people like Elon Musk and Stephen Hawking have warned us about. 

Artificial intelligence Stephen hawking and Elon Musk

But as of now, ASI is just a hypothetical concept. So should you worry about ASI rising as a threat to humans in near future?

Personally, I actually don’t see humans, creating even AGI systems that can rival/surpass a human brain in our lifetime.  But this is just my personal opinion. The thing is, an AGI system requires a framework or a set of algorithms that can encode and learn “common sense” and we haven’t seen much success in that department for decades. Even if we were to make progress there, it’s not like we would jump straight to human-level AGI systems. But rather first we will create systems that can demonstrate rat-level intelligence and then try to build systems for cat-level intelligence and then bit by bit go all the way to human-level intelligence. 

Mouses to Human level Artificial Intelligence

This is a long journey requiring countless innovations and major breakthroughs in the field of AI.

SUMMARY:

ANI systems are successfully being used in multiple industries and it’s being widely adopted at an unprecedented pace. This trend will continue in the upcoming years, AI would continue to evolve and you would see AI taking up tasks and jobs that some of us are doing right now. 

Summary

But there are still concerns about the democratization of AI which need to be addressed, like how it deals with bias in the real world.

Also, AI in the future will be heavily used in weaponry and to track all aspects of your life, but all those systems will be dictated and controlled by people. So AI itself is not something to be feared, the greatest threat to humanity is not AI, it is and always has been humans themselves.

With this we conclude level 1, in the next episode and Level 2 of the series, I’ll go over the history of AI and we’ll understand how we came to machine learning and deep learning. It’s going to be a very interesting post where we will dive deeper into AI.

In case you have any questions, please feel free to ask in the comment section and share the post with your colleagues if you have found it useful. 

Make sure to Subscribe to Bleed AI YouTube channel to be notified when new videos are released.