AI is Smart. People are Wise.

Having rounded the corner on Father’s Day, a few reflections on the nature of wisdom. I find myself continually thinking about how to raise conscientious and grounded children, in a world where the very foundations of identity seem to shift daily. I think about the conversations I want them to be having with themselves, as they interrogate the world around them, and prepare themselves for the future. How do you teach wisdom? You don’t. You set the foundations, invest the time, and hope for the best.

Within all the shuffle around AI, all the whispers and shouts, we're missing a larger opportunity to have a conversation about our human needs and values. About which attributes we choose to anoint and nurture, and which we let recede as atavistic relics of an earlier age. Given that light, it's interesting to note that we don't talk about "Artificial Wisdom" in the same way that we talk about "Artificial Intelligence". Why is that? Perhaps artificial wisdom is a cognitive leap too far for even the most ardent futurists - a leap that's simply incompatible with our notion of what AI is and should be. While it could merely be a semantic difference, maybe, just maybe, wisdom is one of the true moats of humanity in a world under siege by technology. If AI is smart, and people are wise, what's the difference, and will it matter to future generations?

While we're speaking specifically to wisdom in this discussion, much the same could be said of other essential human characteristics. I'm neither well-equipped nor well-qualified enough to properly explore the metaphysical digressions at play here, but instinctively, I believe most people recognize the difference between writing a poem and being a poet; between drawing a picture and being an artist; or in our case, between dispensing wise advice, and being wise. There is something different - but what? I'll offer two key considerations - time and consequences. Let's suppose that AI can effectively simulate the attribute of wisdom, by virtue of providing wise counsel. That does not mean that it is wise. It is lazy thinking to simply state that a machine cannot be wise, because a machine is not human, and it therefore does not have the emotions or qualities we typically associate with humanity. Deeper, underneath all of this, we still have to ask - why? Why can't wisdom be replicated in silicon and code?

I suppose then to answer this question, we'll need to define what wisdom is. For our sake here, I'll define it here as a quality of highly refined judgement. That is, judgement that has been suffused with intelligence and experience. And because experience comes only with time, it means that gaining wisdom generally requires the investment of time. One of the key differentiators of AI-generated content is the lack of an investment of time. If wisdom requires time, can an AI ever truly be wise? One could argue that if humans can further our pursuit of wisdom by leveraging the time of others – the experiences of our friends, and of those recorded in books, art, and music over the centuries – then so too can AI. I’d counter that while AI can benefit from this knowledge, it does so without the investment of time required make it part of one’s identity.

Other, more idiomatic definitions of wisdom may help better contextualize it. Wisdom is the insight gained by having walked through many one-way doors, each of which has irreversible consequences. And while AI might be able to generate wise advice, there is nothing irreversible in its outputs, no consequences to its decisions; therefore, while it can replicate wisdom, the character trait of wisdom is not available to it. Or another – that wisdom is pattern matching, writ large. That is, wisdom adjusts our aperture of cognition to build mental models for how the world works, and how it should work – it distills knowledge into replicable patterns. In this way, wisdom is the application of patterns and principles acquired with time, but which stand outside of time. Wisdom implies life lessons, acquired with the only currency of ours that ever really matters – our time.

Let's also posit that wisdom is a relative concept, not an absolute one. That's to say, it is conditioned by context, culture, and interpretation, though within those constructs it tends to hew toward some universal truths. This means that what is wise advice to some may be meaningless to others - it is up to the dispenser of wisdom to be able to recognize the situation enough to render proper judgement, and the receiver, to reflect upon the validity of the advice. Intelligence can, relatively speaking, be interpreted more objectively. It can be gauged through tests and benchmarks and observations. Wisdom must be viewed through the lens of normed expectations – it is a handshake between parties, an agreement on what constitutes the best way to navigate toward a life well-lived.

But we are again splitting hairs, it may seem, when we talk about wise advice vs. the attribute of wisdom. Or are we? Is there a material difference between giving wise advice and being a wise person? That is, can wise advice come from a fool, and can a wise person offer foolish advice? I'd offer that the answer to both questions is a resounding yes. Which makes our discussion of AI that much more interesting. In more succinct framing, if an AI can effectively dispense wise advice, does that make it wise? AI is nothing if not a pattern-matching impresario. If this alone is a measure of wisdom, perhaps we ought to grant the machines their due. But patterns and principles alone do not make wisdom. Without that prima facie investment of time, and the consequences of that time, these outputs may seem wise, but don’t constitute true wisdom. Like many human attributes, the state of wisdom is a process, not an output.

Wisdom implies life lessons, acquired with the only currency of ours that ever really matters – our time.

A thought exercise. Imagine a precocious six-year-old child hiding behind an old oak tree. The child is holding a book of quotes from the great philosophers of history. You've heard that this is a tree of great wisdom - that is offers the answers to life's great questions. You walk up to the tree, and you ask it a question. The child flips to a page in the book and recites a quote at random from behind the tree. It certainly sounds wise to you, either because the random selection actually holds relevance, because you've primed yourself to expect wisdom, or because the quote is so vague that you assume you're missing something. In either case, you feel you've been offered wise advice and leave satisfied. Does that make the child wise? Of course not.

AI is not wise. It is the most potent pattern matching tool we have ever devised, and will, on balance, offer wiser advice than the average human. But this is, at best, a statistical mimic of a wise person. For all the characteristics that we imbue AI with, it’s extremely important to remember that it’s just a probability machine, generating probabilistic text, upon which we humans interpolate meaning. And if wisdom is reflective of a certain moment in time, a certain cultural agreement, then AI, trained as it is upon human inputs, will adapt poorly to changing social mores and truths. It is the proverbial wise fool; the court jester dispensing inscrutable truths without the self-awareness necessary to be wise.

If we agree then that AI is not wise, and that people (or at least some of them!) are wise, we can turn to the second, perhaps more important question - if wisdom helps us manage the complex world around us, is wisdom still important in an increasingly fractured world? As the world becomes more fragmented, accelerated, and stochastic, is the value of wisdom diminished? When relative truths have their lifespans cut ever shorter by rapidly accelerating change cycles, does wisdom, in some fashion, suffer? I’d argue that while harder to attain, wisdom is more important now than ever, and fear we are witnessing an erosion of wisdom driven in large part by three things:

• A decrease in real-world interactions in favor of digital ones. As people retreat further into their devices, their opportunity to participate in "Real World Events" diminishes significantly. This is compounded by the readily available simulacra of experience and education that the Internet and AI provides.
• An abdication of moral authority by our elder institutions, which diminishes their relevance at a critical time in history and erodes the trust of younger generations. One needs only to look toward politics to see this, but same can be said of many corporations, media, and higher education.
• The rapid acceleration of the pace of change. The swift obsolescence of life skills by technological and cultural change makes it more difficult to apply learned principles.

What's the antidote? Well, societally, we may find we need to further emphasize the teaching of character and critical thinking. Some will argue that AI deemphasizes the need to memorize facts,  dates, formulas, and definitions. I'd argue that we cannot think critically unless we can think independently, and without having this foundational knowledge, we are unable to see ideas through to their conclusion without distraction. If each thought requires that we reach for a device in order to be fully formed, then we have ceded our independence. So knowing, and understanding, the facts is still important. We need to be prepared to ask questions and be resilient enough to deal with uncomfortable answers. We need to throw ourselves into the physical world, not just the digital one, and remember that we have a body as well as a mind.

Bowdoin College recently announced a $50M donation to found an AI Studies department, joining Colby College in creating a mid-Maine AI race. I see the pendulum swinging, albeit slowly, toward a reappreciation of the liberal arts, and the marriage of science and free thinking that will be required to make the most use of technology in the years ahead. I remain convinced that the great challenges of the AI age will be philosophical ones, and to properly take on these challenges, we must be intellectually prepared and courageous enough to draw firm lines between technological efficiency and technology-induced apathy.

We cannot think critically unless we can think independently

Praise of wisdom is not intended to, and does not, lessen the value or impact of intelligence. Intelligence and wisdom are, indeed, blood brothers of the human condition. And yes, the capabilities that AI brings to the table are truly remarkable. Intelligence brings us novel medical treatments, better technology, and abundant food. Wisdom grants us the insight to properly use, share, and spiritually nourish ourselves from these inventions.

A life of action, reflection, and study grants experience, if we are lucky and aware enough to absorb it. That experience must then be tempered through the lens of time, allowed to be washed over by waves of nostalgia, joy, and despair. Only then, with proper time to reflect, gather perspective, and derive life's patterns, does wisdom emerge like a sprouting seed. Over time, this cultivation of wisdom creates the foundations of what we might call a good life. AI can be intelligent. It can even exercise judgement, to a degree. But it lacks true wisdom, and in the long run, wisdom is the line we should hold in our soul.