What is Artificial Intelligence (AI)?

Are you bored of hearing about AI? If yes then maybe overhype is the reason. Maybe it’s because of how it is used to sell you things. Could it be the way it has influenced recent political outcomes?

If you are bored with it, then I have some bad news; AI is here to stay. In fact, it has barely got going.

What is artificial intelligence?

“Artificial Intelligence (AI) is the part of computer science concerned with designing intelligent computer systems, that is, systems that exhibit characteristics we associate with intelligence in human behaviour – understanding language, learning, reasoning, solving problems, and so on.” – (Barr & Feigenbaum, 1981)

How far has artificial intelligence progressed?

To understand why AI is in its infancy, we have to understand the landscape of AI. We have to consider its domains and their maturity. AI might be something of a catch all term and understanding how advanced it is, means understanding how we might measure such a thing.

What are the levels of AI?

One generally accepted model of AI suggest that there are 7 levels of AI:

Rule Based Systems

They are the most common manifestations of AI in the world today, but is it really AI? They rely heavily on rules given to them by humans and they carry out actions based on those rules.

Context Awareness and Retention

These are algorithms that develop a pool of knowledge about the specific domain they’re applied to. These kinds of algorithms are found in social media applications, chatbots and are the reason you get targeted ads after searching the internet for certain topics.

Domain Specific Expertise

This is about the limit of our current journey into the world of AI. These systems can develop and improve their own expertise in specific areas, sometimes even outpacing humans. A great example of this is Google Deepmind’s AlphaGo, which managed to beat a master Go player by playing the game in a way even he didn’t expect.

Reasoning Machines

Here we start getting a peek into the sci-fi end of things. That’s not to say that reasoning machines aren’t being worked on, but that they’re far from perfect at the moment. At this point, algorithms would have some ability to self-reflect, applying mental states to themselves or others, capable of forming beliefs and understanding the beliefs of others.

Self Aware Systems / Artificial General Intelligence (AGI)

Now we have both feet firmly in the world of science fiction, ethics and philosophy. General intelligence will allow machines to learn without knowing the rules of the game, they will be learning the rules. Machines could begin developing personalities and having them change based on outside stimulus; just like us.

Artificial Super Intelligence (ASI)

We are now in the domain of super intelligence, AI that is more intelligent than humans. With each of the steps above completed, the shackles of our bio-brains would be thrown off and machine intelligence could improve itself at unimaginable speeds.

Singularity and Transcendence

The final step is really a continuation of the previous one. With exponential growth in the very concept of what intelligence is, the potential for humanity becomes unbounded. This is the notion that the path enabled by ASI could lead to a massive expansion in human capability.

Where will AI lead us?

Artificial intelligence is already all around us and continues to develop every day. Here at Atlas, we use artificial intelligence to help with a number of things, including making our data crunching platform Atlas Boost more efficient and effective.

My hope is that it will be used more and more to support human innovation in important areas like tackling the biggest challenges of our time, like climate change, health and reduce global inequalities.

As with all things tech, AI should be thought of as a tool to be used for the progress of the human race. With proper application, and the right moral compass at its centre, AI could take us to incredible places.