Thursday 4 October 2018

Artificial intelligence - what next?

Image source: Google
AI or artificial intelligence is perhaps one of the most heavily researched areas of modern day science. With the tremendous amount of promise it presents, coupled with the ambiguity of the very idea of AI, it is something that both exhilarates and daunts people all around the globe.

Scientific works on AI have been going on for more than half a century now. The term was coined by J McCarthy in 1955. According to McCarthy, every aspect of learning or any other feature of intelligence could, in principle, be so precisely described that a machine  could
be made to simulate it, resulting in machines that could use languages, form abstractions and concepts, solve problems now reserved for humans and improve themselves. To put it simply - building machines that harboured human like intelligence.

Now the image that comes to some of our minds could be one of despair or one of great promise. Imagine our lives surrounded by super bots that would perform tedious works for us on our commands and possibly perform tasks that are quite beyond our capabilities. Wouldn't that make life much easier? On the other hand (and in a much worse scenario) we could imagine a world where machines far superior to us fought us - the humans, their very own creators, for world dominion. In the present situation, neither of the scenarios are happening anytime soon (atleast not until another century, scientists believe). 

As of now, we are still in an infancy stage of realising the full potential and possibility of artificial intelligence.  But it is out there - playing roles in our lives, right from applications such as Google Assistant to Siri and Cortana, search engines with the ability to provide lists of useful information out of millions of GB of data available to superintelligent games and self driving cars to AI systems (developed by a team in Germany, France and US) that could detect cancer with upto 95% accuracy. 

All above-mentioned examples are a type of artificial narrow intelligence (ANI) or type I AI. ANI are the most basic types of AI. They do not have the ability to form memories, nor use past experiences to form decisions. They are programmed for performing specific tasks and have no ability to function beyond that. This is the very line that separates the AI that we presently have from the AI that we aspire to have.

The type II or artificial general intelligence (AGI) is the type of AI that has the ability to mimic human intelligence and or behaviour and is indistinguishable from that of a human. They would not just take what we give them but form their own thoughts and opinions and have awareness that was hitherto unthinkable in a machine. In other words they would be dynamic beings, capable of thinking and learning.

The type III or artificial super intelligence (ASI) is the final stage of AI  where it doesn't just mimic human intelligence but surpasses it and it's something we only speculate about. It is something that is quite hard to conceive and accept as humans who are quite adamant about calling ourselves the most superior beings of the planet.

But let us imagine that we do get to the point where they are not just an idea that could evaporate into thin air but are very real and functioning. What then? What are we supposed to expect or look forward to? 

In truth, that is something only time will tell. As said earlier, it could go both ways - the great or the worst things could come out of it for us. On the positive side, we would have machines not just capable of sharing a big chunk of our tasks, accessing places and  performing tasks that are impossible for us, but also save us lot of time performing enormous tasks in the blink of an eye, and also innovating and contributing towards research and development - helping us find solutions to enormous global  climatic crisis. Doesn't that seem wonderful?

But on the other hand, the very same qualities that make them a blessing could make them our doom. Being faster, smarter and in general better than us in everything, aren't these very machines going to replace us eventually? We could very well meet the same fate the small-scale workers met in the face of industrial revolution or when the agricultural revolution ended with the farmers of England being outed by new farming technologies. 

The truth, a more banal one, is that we are letting machine algorithms take over what were human decisions earlier to some extent today, and more so in the future - we will be letting these machine algorithms decides who should be getting a job, which part of city should get developed, whether you should get a loan and in case of a crime what punishment should you get. It might sound outrageous but that is exactly what is happening. And the worst part would be that we would be in no position to question these decisions. 

"Success in creating effective AI, could be the biggest event in the history of our civilization. Or the worst. We just don't know. So we cannot know if we will be infinitely helped by AI, or ignored by it and side-lined, or conceivably destroyed by it," Stephen Hawking once said.

There is no going back now - the Jeanie is already out of the bottle; what we can do is consider ways to harness its powers.

No comments:

Post a Comment