Artificial Intelligence – What are we really talking about?
Lately, I have been hearing a lot of chatter about the future of artificial intelligence (“AI”) and the implications that developments within this space may have on front-office, middle office, and functional operations. While I am not much of an expert on this subject matter, I have become somewhat enamored by the thought of artificial intelligence, spending quite a bit of time absorbing as much content on this subject as possible. Over the last several months I have rekindled my love for all of my science fiction favorites, in addition, to checking out a few interesting new selections, including Ex Machina, The Singularity, and the anime classic Ghost in the Shell.
Unfortunately, this post is not about my love of sci-fi flicks. It is however, about the reality of artificial intelligence in the workforce, and what this will likely mean for the worker of the 21st century.
To explore this topic we must first answer a couple of rudimentary questions about artificial intelligence, which in the context of the workplace is oft referred to as autonomics or robotic automation.
So what exactly is meant by the term Artificial intelligence?
The AI from the movies, is better known in the biz as general AI. This is the type of AI which is intended to replicate human-like intelligence, movement, emotion, and dare I say it – even consciousness. Given that we are likely centuries away from mastering the human brain and body, I think it is safe to assume that we won’t be “playing god” any time soon. Let’s take manufacturing souls off the table and focus primarily on the type of AI which is developing at quite a rapid clip.
The type of artificial intelligence that is referred to in the context of business automation is called specific AI. This is the type of AI which creates rule base machines capable of learning and adapting behavior to perform a specific task. With specific AI, you should be thinking less Terminator or Johnny 5 (love Short Circuit) and more Watson – the IBM omnipresent computer.
Despite being disappointed that Watson defeated my man Ken Jennings, the future of fairly routine processes, is definitely going the way of Watson. Specific AI, will create learning computers which are capable of performing routine or rules based tasks which require learning to alter or change behavior. It is likely that this type of AI will be more programmatical and less mechanical, although, if you are comparing this to manufacturing processes, specific intelligence may be able to perform some pretty interesting manual processes – although this will require significant advances in the area of robotics and material science.
Imagine smart systems in your home which monitor and record your behavior as you come and go. This “smart” home built around specific intelligence would be able to up the heat or air conditioning without the home owner even having to lift a pinky. I think this type of AI could also be referred to as intelligence for the lazy man. But for corporations who hire a lot of people to do routine work, this type of technology is and will become very economical.
What kind of work would this intelligence do?
Since the intelligence of the next decade will be primarily rule based, intelligent systems are mostly constrained to solving very specific problems. To date, there are a couple of very interesting “intelligent” like applications which we use on a regular basis. For example, the airline industry is built around the auto-pilot. Without the software to essentially guide and fly the plane while you are in the sky, I imagine that the pilots may be putting in a lot more work hours. The fortune thing about flying is that the plane, takes in environmental parameters which trigger a series of decisions. This type of automated system is really the beginnings of “intelligence” v1.0. Think of systems capable of making decisions based on external outputs.
If you think of Watson as the next evolution of AI, you end up with a computer that is capable of processing information faster than the human brain. Essentially Watson is capable of taking unknown inputs and quickly querying data sets to produce a statistical model for answering questions. Again all rule based, but Watson’s primarily logic is based on probabilities and sequencing of questions. This type of AI would be incredibly useful for linear input and output type equations.
When we start talking about specific AI today, we are talking about computer systems which are capable of making decisions. These decisions are not made based solely on programmed “logic” nor are they based solely on probabilistic expectation. Instead these machines would base their decisions on “experience” along with probabilistic expectation. As these machines continue to perform tasks, they will continue to get “smarter” and more efficient at their jobs – similar to the way that human beings become more productive today. This is where modern Ai, what I would call v3.0, is distinguishing itself from previous examples of adaptive and smart technology.
Ultimately later phases of this type of specific AI would start to incorporate new features including mechanical and digital integration to not only perform tasks which are codified but also those which are manual or physical in nature or even those that require some moderate level of intuition.
What are the implications of artificial intelligence?
The future of artificial intelligence will have significant implications to the social, economic, and political fabric of the world and the people / institutions that make it up. Today, while aspirational, we are still quite a ways away from having humanoid robots walking the earth. I would venture to guess, however, that we are on the cusp of focused implementations of artificial intelligence for rudimentary process oriented or rules based work. Think any type of transactional type processing work which may happen at home or within a business. Further down the line, we may see more advanced implementations like autonomous vehicles or fully integrated smart systems – although these types of intelligent systems may require significant upgrades to underlying infrastructure.
Ultimately, artificial intelligence will continue cross over between logical and physical systems via integration, however, the future of humanoids is really a long way away. I do not imagine that we will have walking, talking, humanoid surrogates roaming the earth within my lifetime or even several lifetimes thereafter. I do expect, however, that many of the advancements made in manufacturing automation and intelligence to make their way deeper into the workplace and even our homes – ultimately, making our lives much more convenient. On the other side of this coin, I could be way wrong and we could be looking at a dystopic future similar to that of The Matrix. I guess the point here is that if artificial intelligence can be thought up, it will ultimately become a reality – at least under conditions where it is commercially feasible and ethically bound.
For additional examples of robotics and artificial intelligence please feel free to visit the following hyperlinks: