So often nowadays we hear about Artificial Intelligence and how robots are going to take over the world we are constantly reminded of the fact that as humans we are becoming less and less useful and will soon be superseded by computers who can talk like humans and think like humans. Clearly the way we think about Artificial Intelligence is starting to move into more of an obsession and the average person in the street knows much less about how AI works than perhaps they should. Given the vast amount of progression that has happened over the past few years it would be only natural that we believe the hype that we hear in the newspapers and on the internet.
Artificial Intelligence does 3 key things – it learns from the historical trends of activities that we have performed over history, it then applies reasoning to it, by which I mean a set of rules that determine the outcome, then finally it auto-corrects itself – or to the uninitiated it learns from its errors and uses this information to be more accurate in the future. In a way it very much emulates the behavior of a very small child, it can only perform tasks it has seen before, and when chastised for bad behavior it will remember this and possibly avoid it in the future. Unless the benefit of being bad outweighs the chastising! This is known as Weak AI and it really is the type of AI that is used by Google in its new assistant, Apple in the well known SIRI and also the self driving car. Weak AI has become a commodity technology in such a short period of time and its accuracy is clearly growing, but this is like trying to train a dog to talk, I am sure given enough time it is possible, but the effort in this far outweighs the advantages, simply because it takes too much training. However we have made huge strides forward in the direction of Weak AI like this and it has become very useful in some situations, and the progression will continue and my dog one day will say “hello” to me.
Strong Artificial Intelligence on the other hand is a little more of a worry, its is often referred to as General Artificial Intelligence the word “General” being the slightly more interesting term, because this type of AI is not designed to solve a single problem and based on a set of training information, but it is designed to perform much more general activities such as emulating the activities of a cognitive human being. The test of a true AI engine was originally coined by the famous mathematician Alan Turing back int he 1950’s where he defined that true AI is where a process can be presented with a previously unfamiliar task and is able to navigate to a suitable solution without interference. Essentially this would emulate the cognitive capacity of a human – in other words the ability the actually “think”. This does not exist, and in my lifetime neither will my dog hold a conversation with me – nor will my computer.
“Real Intelligence” – is the ability to have a self sufficient, organic instrument that can think, learn, adapt, have feelings and self awareness enough to be able to propel themselves through the world almost effortlessly. Additionally possess the ability to replicate themselves and better themselves over time though evolution, and as evolution evolves, the process could get faster and better to a point where it would be able to re-invent itself and effectively make itself completely obsolete.
Clearly in the last paragraph I was defining human beings, and our insane trajectory to make ourselves obsolete. I wonder if we really thought about this and if we really contemplate what is going to happen to those 7 billion organic beings on the earth when they have no work to keep them busy, no purpose to keep them interested and no reason to awake in the morning? Why do we need to automate to such degrees – to save money?, sure, to be more efficient?, sure – but the real truth is we just can’t help ourselves. The evolution of life is not just about the individual being born, living, dying to be replaced by a newer, better version – it is also about the species, being formed, existing and eventually dying and being replaced by a newer better version. Or so it seems.
I believe in one thing, above all else in the field of computing, artificial intelligence, machine learning, I believe in the concept of a “Human in the Loop”. Humans are after all a huge resource of skills, and learning, that are often undervalued and mistreated – but used correctly they can be very effective. However because we are “only human”, we are also inaccurate and prone to mistakes. It’s these mistakes and errors that define us and make us what we are, and give us our “Real Intelligence” – plus these mistakes and errors keep us busy and in employment.
Mistakes and errors can of course be circumvented and avoided using AI technology for sure, and processes can be made faster and more efficient – but we need to be careful as this starts to break down the whole basis of our society.
At ENTITYSTREAM– we believe in the CAREFUL application of AI and machine learning WITH a human in EVERY loop and that this is the best of all worlds, we will never use AI to replace the work of a HUMAN, but only to allow them to do things they couldn’t do before.
Robert Haynes is a veteran MDM solutions consultant and product director, he has been involved in the Master Data Management world for over 15 years and helps to implement solutions for some of the worlds largest companies. He heads EntityStream as their CEO, EntityStream is a full function MDM solution that can be deployed in a single desktop or across multiple machines, but more importantly it requires no complex setup process nor lengthy configuration of models and feeds. But it does require humans.