When is the last time you went to a library, pulled out a drawer from the card file and jotted down a Dewey Decimal number so that you could go find where a book was located in the rows and rows of shelves?
Fifty-seven years ago, I was taught all about the Dewey Decimal System and how to use it, because if you didn’t understand it, you could look for days for some book you needed for your homework.
Now, your iPad and Siri can find you almost any subject in about one-billionth of a second. There's no Dewey Decimal system needed. That’s part of the magic of machine learning headed big-time into our lives.
Machine learning is the ability of computers to learn on their own by using algorithms that churn large quantities of data. One goal of machine learning is to automate and prioritize routine decision-making so that you can achieve the very best outcomes sooner than you would if you were making those decisions on your own.
Increased data-processing power, the availability of Big Data, the Internet of Things (IoT) and improvements in algorithms are empowering machine learning’s potential to remake not only business, but perhaps, society as a whole.
Many of today’s business processes are governed by rigid, software-based rules, and are thereby limited in the ability to tackle complex processes. Moreover, these processes tend to require hours of employee time spent on boring, highly repetitive work.
For example, when we look at the business finance arena, we see that millions of work hours are spent each year simply checking invoices and expenses for accuracy. With machine learning algorithms replacing the stringent rules, cognitive computing could reveal valuable patterns and solutions that we humans never knew existed. Meanwhile, employees could likely be reassigned to more strategic and challenging work.
Recent advances in machine-learning are trending toward a future in which a variety of bots based upon self-learning algorithms will act more independently than they do now. They may reach conclusions within certain parameters and then adapt their behaviors to the situation at hand. It is likely that such bots will have a great deal of human interaction, both in terms of the service to, and reliance upon, their human counterparts.
Perhaps we will never see a real Robby the Robot going about taking on all the duties and responsibilities from meal preparation, to tending the garden or driving the transport to defending us against invaders. But be certain that the bots of tomorrow will be more sophisticated than those we have already grown fond of.
Like Siri or Alexa, these next generation bots will not only react to our voices, but be much more intuitive and interactive. These learning assistants will continuously monitor our actions and reactions so as to truly understand us, in the same way parents first monitor their children to develop an understanding of who their child is becoming.
That way, Robby may simply ask us at exactly 21 minutes and 18 seconds after our last cup of coffee if we'd like another cup prepared the way we like it. You see, Robby has learned over the last three years that we ask for another cup of coffee about 22 minutes after our first. Now he has become sufficiently aware to ask us.
While this is a very simplistic representation of machine learning, I doubt seriously it will take three years of observation on Robby’s part. By the time Robby has linked into the IoT and evaluated that historically, based upon Electric Company records, the Krupp’s machine in our kitchen consumes extra electrical wattage every 22 minutes consistent with a brew cycle during the hours of 6 a.m. and 8:30 a.m., he has computed in nanoseconds what would otherwise take a human being a ‘marriage’ to learn.
When you think about it, the ability of machines has set them on a course not unlike our own evolutionary path. While computing machines started appearing on the scene in the early 1960s, they have truly begun to expand beyond "number crunching" and started gaining the ability to speak, listen, see, read, understand and interact with ever-increasing sophistication.
In just the last four years, the error rate in machine-learning, driven image recognition has fallen to near zero. Now, it emulates human capabilities (and without glasses no less, like a lot of us must wear to see the tiny little things).
If you remember my last article, I mentioned that the HAL-9000 was able to read two sets of lips simultaneously from a sideview. With machine recognition progressing as rapidly as it is, it’s likely that some thinking machine (HAL of the future), perhaps IBM's Watson, will be able to perform such HAL-like achievements, even though no human can do that or is ever likely to.
When machine learning matures to the point that it can handle unstructured data and when machine algorithms begin to interact with the IoT more freely, machine learning will ultimately become embedded in all systems, devices, machines and software.
Already, machine-learning applications based on these available technologies are being created. We can only guess at the level of autonomy such machines will develop, but be assured that each futuristic step will have a significant impact on society.
Software and hardware developers are striving to build the next generation of learning machines and bots. They know that most businesses, and many individuals, will be looking for that device and network that gives them access to any and all information they need and want from wherever.
Those same developers also know that the systems they're developing will be hungry for a never-ending diet of data. Learning machines are powered by algorithms that would prefer to binge on an unlimited supply of information, in whatever and as many varied forms and sources as they can consume.
Multiple streams of data, streaming from as many sources as possible is where some technologists of the future believe that blockchain comes into play. While thus far limited in practicality, it could very well be that blockchain does for data what the cloud has done for personal and SMB computing.
Prior to the internet and the cloud, data was stored in little silos.
Some data here, some data there, here a data, there a data, everywhere a data, data.
And the IT managers of those little silos, like the librarians of old who were the keepers of the Dewey Decimal system, could keep you from that book you wanted (or data you needed) simply by putting it in the wrong place when they restocked it.
Blockchain would in a sense put all the books in all the places that learning machines could access. In other words, all the data would be everywhere. None of it would be locked away by IT guys, data-custodians or librarians.
This means that businesses will need to have access to data and provide data access in the years to come. To prepare, big business leaders are already working with trusted advisors who can help them identify and evaluate the richest areas for machine-fueled data insight, improvement and accessibility.
These advisors are building a practice based on their ability to address the cultural changes and management challenges business will need to take advantage of during the Artificial Intelligence evolution.
Many people are afraid of what AI will bring with it. They fear that as machine-learning approaches the level of human beings, humans are destined to be overcome by "the rise of the machines." In a machine world, we will be like the guy who made buggy whips (there are not too many buggy whip makers out there today are there).
Of course, many tend to fear what they don’t understand. They're afraid of the unknown, just like setting foot in that huge library for the very first time. It was scary, all those books, sometimes floors and floors of shelves and shelves with rows and rows of books.
You could get lost among all them. Heck, you might even get locked inside if you were buried too deep within the bowels of the library that you didn’t hear them say they were closing in five minutes.
What’s even worse, you could have not even found the book you were looking for yet just because you had never taken the time to learn those Dewey Decimals.
I’m afraid that AI may indeed hold a fate for some of us that’s even worse than getting locked in the library all night. That's especially so if we don’t now learn as much as we can about the next great system on the horizon.
So, ask yourself, are you ready for the change? Are you taking the steps to become informed about the evolution of AI into your business and career? Or, are you sitting around shuffling the cards of your own personal Dewey Decimal system?
My hope is that my "Future of Humanity in an AI World" series will help inform you, and thereby make you better prepared for the changes on their way.