AI vs. Humans, Part II

A few months ago, in an article entitled “AI vs. humans, part I,” I opined upon the promise and peril of artificial intelligence as it comes to occupy an increasingly consequential role in our economy and culture. I called it “part I” because there are, no doubt, many implications of AI that we haven’t anticipated and don’t yet understand.

One of the points I made was that since machines aren’t human, keeping in mind the differences between the two is vital. The power of AI is rooted in taking vast amounts of data that are incomprehensible to the human brain and turning it into information and, ultimately, knowledge that can be used in automated decision-making. Now that we’ve come to understand that knowledge is the source of wealth — the raw materials to create smartphones and spacecraft were available to us long before we knew what to do with them — that’s a big deal.

AI depends on human knowledge

Without question, artificial intelligence (and the automation it enables) can be of great benefit to humanity as we use its strengths to augment our weaknesses. But knowledge is about as far as AI can take us, and knowledge has its limits. The extent to which we subsume our roles as humans to machines will only make the world more, well, robotic. Wisdom — the ability to reflect, discern, perceive and make sense of complex situations — is vital to civilization, and wisdom is a uniquely human trait.

To cite an example, consider the decision whether or not to employ the services of a robotic pen. I recently saw one in action at a trade show, and I have to admit, it was fun watching the mechanical arm methodically and melodically script a thank-you note in near-perfect handwriting. But it was also a little creepy witnessing the production of what amounts to industrially fabricated sentiment. It made me consider that “robotic pen” is somewhat of an oxymoron.

I don’t think anyone who receives a thank-you note written by a robot will believe for a moment that a human actually created it. Why? For one, it’s too perfect. Using a cursive font won’t fool anyone, and even if the engineers were to program in a mistake or a slip of the pen, there would be something phony about it. Beyond that, the language in the note has to be kept rather generic so that it will apply to anyone who receives it. That, too, will subtly telegraph that it didn’t come from a human hand.

It’s true that someone at some point likely drafted the words the robot was to write (although AI can now also do that in a rudimentary way), but that’s a far cry from an individual sitting down at a desk and taking five or 10 minutes to craft genuine, sincere and personalized thoughts in longhand. The inefficiency of handwriting a note is the point; that’s what makes it special (and it’s why Christmas card signatures stamped out by a printing press have always been just as problematic). Having a machine do the work suggests — declares, really — that the recipient is not worth the time it takes to craft something personal. It’s transactional, not relational. Better to risk sending no note at all than one produced by a machine.

The robotic pen, like the printing press, can efficiently transmit information. But it can’t adroitly tailor that information to navigate a complex relationship dynamic. It can’t choose just the right phrasing or gauge the subtlety with which something needs to be said. It can’t convey the sense of worth implied by someone having taken the time to personally pen a thoughtful message. Granted, this example is much less sophisticated than other, more advanced AI applications, but the track record of much of what we’ve turned over to machines has revealed the same shortcomings, from telephone trees to online help desks, chatbots to the endless spam that clutters up our inboxes. To shamelessly crib a ”Power Branding” principle, just because it can be automated doesn’t mean it should be.

Creativity is where humans shine

The Industrial Revolution that began in the 19th century has given way to the information revolution, with factory workers increasingly being replaced by knowledge workers. I find it interesting that the Encyclopedia of Management has a two-part definition of “knowledge worker.” It begins by defining them as “those who acquire, manipulate, interpret and apply information.” That’s a task that is increasingly being automated, a development we’re all witnessing and, to some extent, welcoming. But the definition goes on to say they do so “in order to perform multidisciplinary, complex and unpredictable work” and “apply expertise in a variety of areas to solve problems, generate ideas, or create new products and services.” Those are skills that are uniquely human. And that’s where real value lies.

In AI circles there’s talk about “the singularity,” the moment when machines become more intelligent than humans, with each generation of machines then able to create even more intelligent machines and exponentially leave human intelligence in the dust. But that’s based on a narrow definition of intelligence, neglecting the role of emotion, nuance, judgment and, yes, wisdom. AI will get increasingly better at mimicking human characteristics — “aping them” may be a more fitting term — but it will never be human. There’s a reason it’s called “artificial.”

A machine can create art, but it can’t be awed by it. The moment you automate a relationship it ceases to be a relationship. And relationships are the stuff of life.