The calendar is about to enter a new decade. The 2010s are about behind us and, in a week, 2020 will be on hand.
With the start of the new decade, we look ahead to see where artificial intelligence will take us. Basically, it is closer than we think.
Within 5 years, we are going to see widespread super-intelligence. These AI systems will be operating at a level far beyond what humans have imagined.
Before getting into the details, please bear in mind I am not referring to a super general intelligence. This is not a proclamation on AGI where computers will be thinking like humans. Nor is it within the bounds of this article to speculate on that.
Instead, what is being referred to here are narrow intelligence systems. Artificial narrow intelligence is a super-intelligence that is rather dumb. We have seen instances of it with games. An ANI was able to beat the best Go player in the world and even pull off a move humans never even considered. Go is thought to be the toughest game on the planet. Yet, the intelligence that beat the best cannot tell the difference between a cat and a dog. In contrast, any child knows the answer to this.
The last hundred years saw remarkable economic and societal progress due to specialization. I guess you could say this really blossomed with Henry Ford and the assembly line. This ushered in a new era of people doing specific tasks or having specialized knowledge.
Unfortunately, this is right up artificial intelligence's alley, at least what we have today. AI is very poor at general aptitude stuff but when it comes to specialization, it dwarfs human ability.
This is what we are going to see expand over the next few years. Already we are seeing, although there is a bit of debate. AI systems outperform radiologists. In fact, automation is trumping any type of human scan recognition in the medical field. Elon Musk claims that his cars, with even minimal AI features, are safer than those driven exclusively by humans. Money managers on Wall Street have a difficult time beating the S&P yet AI systems are dwarfing their returns.
Even if these examples are not quite true yet, the is no doubt they will be within a year or two. Thus, while we still lack the ability to create AGI systems, we can create ANI that are far superior than humans at specific tasks. Once we see one developed, it simply is a matter of creating another. We see this happening across all industries where millions (billions?) of algorithms are being written.
By the middle of the next decade, we can expect a series of AI systems to be operating at what we consider to be a super-intelligence level. None of them will be a threat to humanity since, outside of their specific task, they will be as useful as a wet mop. However, when strung together, they will operate in a manner that takes our intelligence level and sends it to the moon. We already have seen a bit of this in the financial arena where humans were replaced with computers.
The impact this will make is not yet known. We are about to embark upon the great period of unleashing power we have ever seen. Our economy, along with other societal factors, are going to be transformed, something unseen since the Industrial Revolution.
One of the big factors I see today is the growth of the open source and decentralized communities. As these realms expand, more impact will be felt from them. This is going to alter the power structure in many ways. It is also going to offer opportunities that were never available to the general population before.
However, there will be bumps in the road. This is not going to be a smooth ride. There are those who are negatively being impacted also. This is where we need to step things up from a discussion level.
Unfortunately, few seem to be aware of how close this really is. Most are thinking in terms of an AGI and when that will come. The reality is, a hivemind of ANIs is much closer than people think.
If you found this article informative, please give an upvote and resteem.
