Reports on AI and education are full of many a sleight of hand. The recent report from the Brookings Institute is no different. The best I can say of it is: their hearts are in the right place. Yet despite expressing some strong concerns at points, they betray their true position early on. They report, “We find that AI has the potential to benefit or hinder students, depending on how it is used.” Not only does this reveal that they hold to the waning belief that technology is a neutral object in the world (without any power to influence the world by merely existing), this sort of neutrality — it all depends on how you use it — is often the telltale sign of what’s to come: a series of well-meaning but empty recommendations.
After reading these reports over the years, not merely reports on AI, but on any emerging technology, one eventually has the feeling of deja vu. The patterns of thought begin to reveal a boilerplate state of mind, susceptible to putting forward claims without having thought them out with any real reflection. These ideas want reporting. Hence, I have found a few moves that repeat across these reports. What follows is meant as a reference point for some of them.
Sleight of Hand Iº
Anything can be done "well"
Consider these, the two upshots from their report:
AI-enriched learning. Well-designed AI tools and platforms can offer students a number of learning benefits if deployed as a part of an overall, pedagogically sound approach.
AI-diminished learning. Overreliance on AI tools and platforms can put children and youth’s fundamental learning capacity at risk. These risks can impact students’ capacity to learn, their social and emotional well-being, their trusting relationships with teachers and peers, and their safety and privacy.
What do these mean?
Consider the language of the first upshot, that well-designed tools will reap benefits. What, pray tell, is the support for the view that there is such a thing as a well-designed AI tool?
Turn your gaze to Athens, to the briefest of admonitions. Visitors and seekers of wisdom at the Temple of Apollo would behold it high above them, etched into the architrave, as they passed the chilly marble pillars, once ponderously shipped from the Pirgon quarry or some other hole in the ground.
μηδὲν ἄγαν. Nothing in excess.
Whether or not we agree with the sentiment, we can agree this translation means something very much different than another common version: everything in moderation. One might consider these as amounting to the same thing. Yet they differ in at least one respect: the scope of actions which they lay claim over. The first translation of the admonition, on the one hand, does not suggest that every possible thing one might attempt should be tried. There might, for instance, be some things I should never attempt, moderately or otherwise. The second version, on the other hand, seems to suggest that we might try everything, but the trick is doing any of those many things all in moderation. It’s the difference between “I’m not saying do it, but if you do, do it in moderation.” Vs. “Anything goes, just moderately, mind you.”
But we would do well to remember our Greek tragedies.
First explorer, seeking wisdom from the oracles: So, should I run at the Siren in order to get her all scared and apologetic?
Version Iº: Nothing in excess. (The explorer leaves, puzzling it out; When the Siren shows her teeth, he sails away, but not to the excessive limits of the world.)
Second explorer, seeking wisdom from the oracles: So, should I run at the Siren in order to get her all scared and apologetic?
Version 2º: Everything in moderation. (The explorer, when the Siren strolls into view, charges —moderately — and threatens her with his sword — moderately. Later, the Siren adds a moderate dash of salt upon her man-steak.)
Some things can be done, but never well.
Sleight of Hand IIº
The confident tautology
Yet another trick is spied up their sleeves when we pause to consider the emptiness of their upshots in their conditional forms. We will recall a classic example: if there’s smoke, there’s fire. What sort of conditional do we find here?
For the sake of clarity, and since I’m in the translating mood, here’s my excessively wooden, protestantly literal attempt to convert Brookings-speech into the vernacular:
If you use it badly, our very fancy thinking caps tell us that bad stuff will come out.
If, however, you use it goodly1, then we discovered after hours of reflection that good things will happen!
Well’s contribution is perhaps to clarify away from ill-designed tools, yet it also has the power to smuggle in the conclusion before its time. After all, when the condition is to do something well, does the conclusion, where all is well, at all surprise us? Declared with authority and the air of serious reflection, it’s still tautological, a sort of ‘the thing you should do is the thing you should do,’ or in conditional form, If things are done so that benefits accrue, then things are done so that benefits accrue.
As a simple test, whenever one encounters something in the shape of ‘if it’s well-used, well then buckle up for all sorts of good things to follow,’ see if the outcome feels as viable after omitting “well” from the claim.
Example: If M. Night Shyamalon’s The Last Airbender is watched well, the genius behind the camera, the care of the script writers, and the acting prowess of the actors will shine forth.
The idea behind this can be innocent enough. It wants to gesture towards many a venerable idea: Put good in; get good out. Like is drawn to like. Like produces like. They hope to glory in these ancient dogmas, flowering in this, the unlikeliest field (of research). The goddess Phoebe shines upon their labors (=research findings) as they march in parade from the temple of Apollo.
Sleight of Hand IIIº
Revolutionary change disguised beneath equivocal continuity
Here the idea is that we once placed books into the student’s hands, and thus we educated them with great success. We are involved now in the same project of education, just with AI placed their disposal.2 In replacing one tool with another, the outcome remains as it was. The tool used does not have a hand in (re)shaping the outcome.
In 2010, the world cup was hosted in South Africa. Don’t ask me for any of the stats. Not that any of the players themselves would enjoy reading their own from the tournament. It was not a great year. It was the year of the Adidas Jabulani, a soccer ball widely repudiated by the athletes. Rather than receiving a good spin, its trajectory tended to go all knuckleball.3
The introduction of something new — by all accounts as inconsequential as a soccer ball stitched in a new way — had an effect on how the game was played; it messed with the player’s expectations; their strategies had to change, and with it the nature of the game.
Introducing a differently-sewn soccer ball did not go unnoticed. Imagine introducing an oval shaped ball to the game, or a ball in the shape of a pyramid, or one filled with lead. At what point will the game have to change (if it’s not abandoned entirely)?
Imagine a time when the book was the dominant tool of education: Italy in the 1500s, or Paris in the 1300s (even Germany in the 1830s, when the book’s dominance was under strain). If we were to introduce an AI thingumabob to the provost of, say, the University of Bologna or to a professor of scripture at the University of Wittenberg, or to that one grouchy philologist at the University of Basel with the bristling mustache4 — you would witness all manner of reactions. What they would have in common would be this: to a man, it would seem as revolutionary to the field of education. Utterly transmogrifying.5 What it would be mean to educate would no longer be the same. What passed as teaching before will no longer obtain.
There is the idea that we may exchange techniques and assume everything else remains basically the same. If we have been using books, say, as the cornerstone tool for education, we can replace or augment that situation by introducing AI-tools, and there’s no need to double check if anything has been radically altered.
One of the key insights from Martin Heidegger and others is that the character of the tool informs the arena in which the tool is being wielded. The tool shapes the space, even before we have used it.
With the introduction of a new tool, we must be on the look out for equivocations. We need to consider if education, as understood under the regime of the book, will be replaced with something that, properly speaking, awaits a different title.
There comes into the world a new tool, largely untested and brimming with the power to alter deeply the space into which it is introduced, and if we continue to believe we are playing with the same soccer ball we’re accustomed to, we will find ourselves at a loss to explain all the stubbed toes and frustrated athletes.6
Footnotes
-
I do not apologize for the metaplasmic proparaleptic form of “good.” I said it would be grossly literal. ↩
-
Ignore the fact, they beg, that AI cannot be placed in someone’s hands; something else must be through which AI is seen. ↩
-
Research after the fact seems to have verified this. ↩
-
The prof. with the mustache, not the university. ↩
-
Their wish for such a revolutionary change in the university system — designs to disrupt, overthrow, and burn, or to reform, strengthen, and expand — would largely determine whether they asked you how to log in or threw you out the highest window. ↩
-
Earlier versions of this post were being edited while they were being posted. I am deciding upon a way to preview posts without having to expose them to your patient attention. My project in self-ownership (and its responsibilities) continues… ↩