Main Street News

What, essentially, is intelligence? The longer we spend on this earth, the more nuanced our definitions become as they incorporate qualities such as intellect, emotion, and insightfulness, some of which can be cultivated while others seem innate, native. For as long as students have attended school, formal education has attempted to inculcate the markers of intelligence in young people so that they can attain success in life, but as nuanced as our definitions of a successful life can be, so too are the markers of intelligence that we transmit to the next generation.

Etymology

Let’s start with the word itself. Having passed through Old French and 700 years of history on its way to modern English, the word “intelligence,” originally from the Latin for “between” (inter-) and “to choose, pick out, read” (-legere), literally describes one who is able to choose correctly between words or ideas that may have similar shades of meaning. 

As one applied arts teacher I know puts it, a word like “tolerance” means one thing in a social context but something entirely different in a metalwork shop. The College Board establishment continues to prioritize that understanding of vocabulary in its SAT and current Advanced Placement English exams that demand the most finely tuned sensitivity to these shades of meaning.   

So it is natural to value skill with words and concepts as a signifier of intelligence (my high school Latin teacher used to say it was the quality that separated us from animals). Teach someone to put the right words in the right order, and you will have good writing and good thinking, to paraphrase Samuel Taylor Coleridge. In a post-Netscape era, when the Internet provides facts and information well beyond the scope of an individual human’s memory, this semantic aptitude trumps the kind of intelligence that wins at Jeopardy, Ken Jennings’s success notwithstanding; even he lost to Watson (a fact I had to confirm with a Google search).

Fakery?

But what comes to mind when we add the word “artificial” in front of “intelligence”? Insinuations of “art” and even “artifice” shade the meaning, suggesting a certain amount of fakery and disingenuousness. Indeed, there is a layer of subterfuge involved when artificial intelligence masquerades as a human being, recreating our speech patterns. We are led to believe that we are actually talking to someone. When we talk to Siri, “she” “talks” back, involving neither an actual person nor an actual voice. 

Children are often deft testers of these technologies because they will talk to artificial intelligence without the conventional posture that adults have come to assume in relation to robots. They will ask her to tell them a joke, pose nonsensical questions, and even inquire about Siri’s origins, relationships, and emotional life. This all provides Apple with more fodder to generate appropriate responses to questions like, “Siri, what’s your last name?” With each use, the dissimulation becomes better by learning the signifiers of a genuine human response.   

Predicting… what?

The large language models that fuel artificial intelligence essentially use massive amounts of text along with evolving logarithms to predict the word that will most likely occur next in a sentence. More and more, as Gmail’s predictive typing demonstrates, it is often correct, especially in the antiseptic white-collar rhetoric of the business world. Since 2006 – almost twenty years now – Reuters has used artificial intelligence to produce financial news stories, which prompts the question, “Why should we teach financiers to write?”

Writing, of course, is not grammar, but when too much emphasis is placed on the finished piece of writing and not the evolution of thought that is experienced by the writer, we have missed the point. AI can produce competent writing, but it has yet to produce thoughtful, soulful writing. Rhetorical styles are as unique as fingerprints, often characterized by ungrammatical constructions, euphemisms, and speech patterns that are grafted onto us through our upbringing and education. They are signifiers of personhood behind someone’s words, and good writing instruction works to coax that voice onto the page. 

So to an extent, the ability to parse meaning and to detect artifice is now the challenge of education. Just as Neo notices a glitch in computer code that reveals the Matrix to him, educated people have a finely-tuned ability to spot when writing is not real. 

AI faces an even greater challenge in creating artificial voices, faces, and interactive technologies, which most people, regardless of education level, can detect. And yet, in order to improve our ability to detect impersonations, we are learning more about the signifiers that betray them, which, in turn, AI will avoid. Presently, we can look at the state-of-the-art, AI-generated face and just know, intuitively, that it is fake. There is no soul in there. Developers will study just what happens on a visceral level when we become aware of a soul – or whatever you prefer to call it – and before long, they will be able to recreate it.

Human vs. artificial intelligence

If you accept that chain of events as inevitable, what, then, is the role of education in preparing for the day when we cannot detect artificial intelligence from real, human intelligence? The answer is no less than a redefinition of intelligence itself. Verbal and nonverbal acumen will be replaced by those qualities that define humanity: our ability to think creatively and empathetically. But what do those mean, and what do they look like in a classroom?

AI can think expansively, but thinking creatively involves digressions, distractions, and the alchemy of sleep and inspiration that occurs within the human brain. We have the ability to think about an idea, discuss it, walk with it, and come back to it a day later with new eyes that can produce profound insights. 

Examining literature

I am reminded of a course I took that paired Shakespeare and literature of the Vietnam War. The essays would take days to conceive, contrasting Henry IV with Robert McNamara through a process that benefited from a few day’s distillation as I went for bike rides or simply procrastinated. The insights were profound, and lasting, and my mind was changed by the work. 

I am also reminded of a student in one of my poetry classes who selected “The Daffodils” by William Wordsworth as a poem to memorize for an assignment. Reciting it aloud in his kitchen, his grandmother chimed in from the next room, falling into the rhythm of the lines she had memorized decades earlier. Artificial Intelligence can read poetry on command with consummate accuracy. But poems that mean something to us etch their cadences onto our brains in ways we can’t explain. The “austere and lonely offices” of the father in Robert Hayden’s “Those Winter Sundays” generates a complex image of fatherhood and duty that shapes my relationship with my family and warms me in a way no thermometer cannot measure. 

Poetry and art lean heavily on empathy to do their work, and in an era when technology has the potential to do so much good as well as evil, the best science will lead with the heart, too. The myth of the scientific process’s sterile objectivity has yielded to the humanism of approaches such as the Stanford Design Process, where interviews and rigorous compassion generate ideas and solutions to problems of the human experience. Artificial intelligence could support this process as a tool to kickstart ideation, but it cannot outperform humans when it comes to considering the impact of design on the human experience. 

It boils down to this question

Next-generation science practice places students at the center of science instruction instead of an experiment, and their ability to understand, rationalize, and ideate solutions provides a learning model that prioritizes real thought and emotion over fact recollection and regurgitation.

Ideologically and philosophically, this boils down to one question: what is human intelligence able to do that artificial intelligence cannot? The answer lies in knowing how AI technology works, and just as importantly, in knowing how our brains work in ways that are different from that. 

We can know the former, and we should educate the next generation about how computers recreate – imitate, really – forms of intelligence while helping them to harness that power. We are starting to understand the latter, thanks to fMRI scans, but as much as we know about the electrical impulses that accompany our learning, we are just scratching the surface of how electricity, blood, and chemicals commingle to influence all aspects of human potential. We know still less about whether those biological elements are the sum total of what defines us as learning, growing, humans. I, for one, suspect there is much more to learn. •