Let's Put AI in Perspective

Artificial intelligence coverage these days falls into two camps: breathless rhapsodizing on how AI will make life easier, work more efficient and create the professional equivalent of unicorns and rainbows; or equally strident takes about how AI will make Skynet real and bring Armageddon down upon us. While both arguments have a basis in truth, “basis” is not the same as reality. So, may I suggest that we all take a collective breath and consider the facts?

Data Isn’t Knowledge

First, let’s talk about what AI is — which, by definition, also introduces what it isn’t. Let’s begin with the name. Yes, it is artificial. It is made of silicon, and is therefore not alive or sentient in any way. As for the intelligence part, the term is an unfortunate choice.In my mind, “AI” should actually be an abbreviation for “Augmented Insight,” not “Artificial Intelligence.” Let me explain.

The Webster Dictionary defines intelligence as “the ability to learn or understand or to deal with new or trying situations.” AI as we know it today fails at all of those things. The same dictionary defines learning as “the ability to gain knowledge or understanding of or skill in by study, instruction or experience.” An instance of AI can’t gain knowledge; it can gain data and information, but knowledge? Knowledge is defined as “being aware of something.” AI instances aren’t aware, because they lack the senses that we have that allow us to comprehend our world through synthesis, a process that we don’t even understand. In fact, I maintain that AI can’t learn, but it can be taught. There’s a difference, and it is fundamental.

Polanyi’s Paradox in Action

AI technology learns (to the extent that it can) because humans make information available to it.In other words, humans teach AI what it needs to know to do things. But humans are biased — what’s important to one person is immaterial to another. So, why include that particular data set in the pantheon of what I want the AI to have access to? Furthermore, while AI has access to vast stores of data, and the ability to relate that data, it doesn’t come close to understanding it.

Polanyi’s Paradox tells us that ”we know more than we can tell.” Want evidence? Ask yourself this question: How did you get up this morning? If you’re like most people and you slept in a bed, your answer will be something like, “Well, I sat up, and then I…”

Stop there. How did you sit up? How did you swing your legs over the side? How did you stand? How did you walk to the bathroom? If you think about this for mere seconds, you’ll conclude that you have no idea how you did those things. We know more than we can say.

And now you want to infuse an AI entity with all it needs to think and act like a human. How would you suggest we do that, when we don’t even know how WE do the things we do?

I am not a closet Luddite. I’ve spent the last 40 years writing books about technology and working as a consulting analyst to the greater tech industry. But after spending all that time around technology, whether we’re talking about robotics, 5G, optical networking or AI, I’ve come to look at all of through a thick and focused lens of skepticism.

Related Article: Hire for the Full Spectrum of Competencies

The Wheel, the Screwdriver and AI

Is AI a game-changer? Absolutely. But so were the lever and fulcrum combination, the pulley, the wheel and the screwdriver. They’re tools. They don’t replace humans; they augment human capability. What AI gives us, more than anything else, is velocity. Yes, it can write a five-page paper on any topic you like, and the writing will be exquisite. When I recently asked it to write me a short paper on a current topic (and I did this several times, by the way), it did so, beautifully. As a professional writer, I was a bit chagrined when it took 90 seconds to create something that would have taken me a few hours. But I was less chagrined when I discovered that every quote, every bibliographic entry and every supporting fact in the paper were fictitious. It made them up. That’s not helpful. Yes, I can use the words that it strung together in the essay, but now I have to go through and validate every fact upon which the paper is based. In other words, I don’t trust it.

So: back to my original contention. Take a breath. Is AI helpful? Yes. It can analyze vast volumes of data in a short timeframe that would take a human hours, days or weeks to complete. Does it replace humans? Absolutely not. It augments us, it provides useful insight and it saves us time.

Learning Opportunities

Conference