"Across the country, a small but growing number of educators are experimenting with oral exams to circumvent the temptations presented by powerful artificial intelligence platforms such as ChatGPT."

"Such tools can be used to cheat on take-home exams or essays and to complete all manner of assignments, part of a broader phenomenon known as 'cognitive off-loading.' [One teacher] tells her students that using AI is like bringing a forklift to the gym when your goal is to build muscle. 'The classroom is a gymnasium, and I am your personal trainer,' she explains. “I want you to lift the weights.'"

I'm reading "To AI-proof exams, professors turn to the oldest technique of all/A growing number of educators are finding that oral exams allow them to test their students’ learning without the benefit of AI platforms such as ChatGPT" (WaPo)(gift link).

I stare the nerve-wracked student in the face and say "Is using AI is like bringing a forklift to the gym when your goal is to build muscle?"

The student, knowing his grade in my "Logic and Language" course depends on how fluently and sensibly he responds to that prompt, answers:

"Professor, I think the analogy is seductive but ultimately misleading, and examining it carefully reveals something important about the structure of teleological language and the way we smuggle normative assumptions into metaphorical claims. When someone says that using AI is 'like bringing a forklift to the gym when your goal is to build muscle,' the metaphor only works if we’ve already stipulated that the exclusive, intrinsic purpose of 'going to the gym' is the autotelic activity of unaided muscular exertion. In other words, the analogy presupposes a purely expressive or ascetic concept of training: the value lies in the struggle itself, not in any extrinsic end. Under that stipulation, yes, the forklift necessarily frustrates the telos, because the telos has been defined as resistance against unassisted human tissue. But notice how question-begging that stipulation is. In ordinary language, when we say 'I want to build muscle,' the phrase is almost always instrumental rather than purely expressive. The speaker typically means 'I want the morphological and functional adaptations we call 'muscle hypertrophy' because they serve further ends (health, aesthetics, athletic performance, social signaling, etc.).' Once the goal is construed instrumentally, the normative force of the analogy collapses. A forklift would then be evaluated exactly like any other technology: Does it produce the desired physiological outcome more efficiently, safely, or reliably than the alternatives? The answer is often yes, in which case refusing the forklift would be a category error, like refusing to use a pulley system when building a cathedral because 'real stonemasons lift the blocks with their own hands.' So the metaphor covertly equivocates on the concept of 'building muscle.' In one sense (the expressive, autotelic sense), AI is indeed a forklift: it removes the very resistance that constitutes the activity’s internal good. In the other sense (the ordinary, instrumental sense in which almost everyone uses the phrase), AI is more like a well-designed training program, forced reps, or even anabolic steroids: an external means that accelerates progress toward the actual end-in-view.The deeper philosophical point, I think, is that technological mediation doesn’t inherently alienate us from a practice unless we have already romanticized the practice as requiring unmediated struggle. That romanticization is itself a historical and ideological choice, not a logical necessity. Aristotle could still achieve eudaimonia while dictating to a scribe; the medieval scholar could pursue contemplation while using an astrolabe. The tool doesn’t negate the human excellence unless the excellence has been artificially defined as 'doing it without the tool.' So to answer directly: No, using AI is not like bringing a forklift to the gym, because most of us are not in the gym for the sake of lifting weights as an end in itself. We are there to transform ourselves in certain ways, and the intelligent use of tools (including cognitive tools) has always been part of how human beings pursue self-transformation. The real question is whether we can retain critical agency in choosing when the tool serves the end and when it quietly substitutes a different end altogether. That’s a question of practical wisdom, not a logical contradiction inherent in the technology itself. Thank you."