Facts and Myths about Artificial Intelligence: Weak AI, Strong AI & Super AI

A few days back, I came across an article that was convincing people that Artificial Intelligence or AI cannot progress beyond what it is now. It totally ignored the future. But we all know that research on the subject is being carried out more aggressively than any other time in the past. We already have AI based code and machines etc. that will keep on improving. Let’s check out some facts about Artificial Intelligence while also dismissing some myths about it.


Facts about Artificial Intelligence – Conscience Doesn’t Matter

Some articles on the web say machines cannot have a conscience. They go on to say that conscience cannot be measured and cannot be considered scientific. But we know that it is there inside everyone. People have conscience whether or not science considers it as a fact. Science cannot even measure brain or neural networks, but it has to acknowledge them.

With that as base among facts about artificial intelligence, I’ll try to explain the technology and things associated with it.

The main argument in the article was that since conscience cannot be programmed into machines, the dumb AI is the only AI and totally ignores the problems other researchers see with Super AI. There is indeed a school of thought where people follow this theory of conscience and self-awareness. They believe the current AI is not AI at all but just automation just because conscience and consciousness are missing from machines.

But we already know AI is more of enacting human brain… much more than Automation… as it involves thinking on behalf of computers. Thinking on behalf of computers/machines would mean NIL involvement of operators. I’ll cover automation vs. artificial intelligence in a separate section in this article. First, let’s check out the types of Artificial Intelligence.

Read: Debate on artificial intelligence: Will playing God help us?

Types of Artificial Intelligence

You probably already know about Artificial Intelligence. It is categorized into many types. For convenience, we’ll just place them in three broad categories that beginners use.

  1. Weak Intelligence
  2. Strong AI
  3. Super Artificial Intelligence

Weak AI has been around for a long time. It is just a machine that works for which it is designed. Take the vehicle speed controller as an example. The speed controller measures the speed of vehicles and keeps it in control. There is no intelligence there, but it sure removes the need of a human to keep watching the speed of the vehicle.

Another weak AI example is the game of chess. If you are playing chess with a machine, it looks as if it is thinking. In reality, it is parsing all the possible moves and the casualties associated with each move. That is not slow but still sends a signal that the machine is thinking.

In short, where the machines are made to perform few acts based totally on the coding, it is weak AI. You don’t expect real intelligence there.

Strong AI is one where the machines can play the part of understanding. In weak AI, the understanding part is reserved for humans who code or use the machines. Strong AI can understand and take actions based on their actions.

If you remember TAY, the Microsoft chat bot, it was fed with crap all day by humans and became racist in a day. So far it was weak AI. But if it had understood what information it has to reject and what to collect, it would have been strong AI. That is, ignore people who were feeding it with posts that led to negativity and it understood that it is a bad knowledge that has to be rejected, we could have called it an example of Strong AI.

Super AI is connected to consciousness. This is where the machines know they exist and that they exist for a purpose. We don’t have much of Strong AI at present, and hence we consider Super AI as non-achievable. Looking at the other side, research is on; experiments are being conducted, and we may soon have a machine that knows what it is and what are its aims. It’s just a matter of time!

Facts about artificial intelligence – Does Automation mean AI?

No. Automation can be said to be a part of artificial intelligence. Automation can be as small as a baggage conveyor belt. In the speed control example above, the machine which is reading speed can be modified a bit to bring the vehicle back to normal speed. Likewise, driverless cars that can decide where to go can steer the vehicle and apply acceleration. So while the thinking part of driverless cars can be said to be Artificial Intelligence, other aspects such as steering and acceleration are automated. Hence it is safe to say that automation is just part of the output of artificial intelligence.

Read: What is Google RankBrain.

Will Artificial Intelligence replace Jobs?

Confusion about artificial intelligence is that it can replace humans in different jobs in future. Technology as a whole has already cut short the job market, making many of us unemployed. If machines can think for themselves and the business houses, they’ll do it at a speed much higher than that of humans. Needless to say, businesses will invest more in artificial intelligence.

However, there is always scope for other fields opening up. With new technology, there come opportunities too.  Just that the requirement or the skill set would be different. If for example, driverless cars can replace driving jobs, we’ll need a skilled workforce that can maintain these cars.

In this post Microsoft explains how Artificial Intelligence will shape our future.

Can Super AI be dangerous?

It may or may not be. It depends on what is fed to the machines. If we look back at the TAY example, it went bad because people fed it negative things. Likewise, if the machines earn consciousness and learn to ignore bad, they will be good. If there is evil working on it, they may definitely start countering humans. Many of the renowned physicists signed a petition to let governments understand the risks of Super AI and to stop funding such projects. But we know that corporate business houses in their pursuit of profits can go to any extent. Even if such projects are stopped by governments, some corporate houses will be more than happy to fund such researches and experiments. It is then that we need to worry.

Read: What is Machine Learning and how it is different from Artificial Intelligence.

The above facts about artificial intelligence are not comprehensive. I will come up with another article in a while where we’ll talk more of myths about artificial intelligence.

Posted by on , in Category General with Tags
Arun Kumar is a Microsoft MVP alumnus, obsessed with technology, especially the Internet. He deals with the multimedia content needs of training and corporate houses. Follow him on Twitter @PowercutIN

Leave a Reply

Your email address will not be published. Required fields are marked *

2 + 4 =