Assessment Was Broken Before AI, Author Argues

This is Part 2 of a three-part series on AI in education. The author begins by reframing a conversation with higher-ed educators who fell silent when he suggested assignments should not have word counts. His point: AI did not break assessment; it exposed that assessment was already broken.

Word counts, page minimums, five-paragraph essays, and unproctored take-home exams were all designed around the assumption that producing text was difficult, so requiring large volumes of it proved effort. AI removed the friction, leaving behind a pile of assignments that were never actually measuring what they claimed to measure. Outcome-only grading—judging work by its final product rather than the thinking behind it—has always pushed students toward the wrong behavior, producing "fine-looking results" for fast writers and good test-takers while missing what actually matters: how a student arrived at an answer, what they revised, what feedback they used, how many attempts it took before something clicked.

The author then addresses the loudest fear: cheating. A national survey of college faculty found that 78% said cheating has increased since generative AI became widely available, and 73% have personally dealt with academic integrity cases involving AI. One Brown professor recently said flatly that he suspects most of his class used AI to cheat. This is where the panic is loudest, the author notes, because "a college degree is supposed to certify that a person can actually do something, and there's a lot riding on that certification staying honest."

Schools have reached for a quick fix: AI detectors. But these tools are themselves broken. A Stanford study tested seven of the most widely used AI detectors and found they flagged 61.3% of essays written by non-native English speakers as AI-generated, compared to a 3.2% false-positive rate for native speakers. Simpler vocabulary and more formulaic sentence structure—completely normal for someone writing in a second language—reads to these tools exactly like a machine. The author calls this outcome "disqualifying," explaining that the "solution" schools reach for first is actually "a machine for punishing English-language learners specifically." Most professors don't rely on these tools in practice; they trust their own read of a student's writing instead, and the author says "that's the right instinct."

Plaigarism belongs in the same conversation, the author argues. It is not a problem AI invented; high-profile plagiarism scandals have occurred in recent years among university presidents, authors, and journalists, none of it requiring generative AI. AI gave an old problem "a faster way to happen and a bigger microphone."

If detection is not the answer, what is? The author proposes "show your work"—a practice math teachers have required for decades. In his own high school coding classes, he lets students use AI to write code on one condition: they must show him every prompt they used, explain why they wrote it that way, how they tweaked it, how they validated the output, and how they made it theirs rather than just accepting it. "That's not a workaround for AI. That's just what 'did you actually learn this' looks like once the answer alone stopped being proof of anything." Documenting the process instead of just the product makes both cheating and plagiarism much harder to hide. The author adds a caveat: in his own schooling, "show your work" often got him in trouble because he did a lot of math in his head, and a teacher who assumed there was one correct method marked him wrong for not showing steps he hadn't taken. If "show your prompts" turns into "there's one approved way to prompt," he warns, "we'll have rebuilt the same trap with different materials. The point isn't to mandate a process. It's to require honesty about whatever process actually happened."

On critical thinking, the author calls the fear "the calculator panic, running the same argument on new hardware." In the 1970s, handheld calculators arrived in classrooms to nearly identical warnings: kids would forget how to do math, number sense would atrophy, a whole generation would grow up mathematically illiterate. Fifty years later, schools still teach arithmetic by hand before calculators, and no one argues schools should have kept them out. The tool did not replace thinking; it changed what thinking got spent on. "Explaining why you wrote a prompt the way you did, evaluating what came back, how you iterated and fine-tuned, and knowing when to trust it and when not to, is itself a critical thinking exercise, arguably a harder one than solving a problem that already has one correct answer waiting in the back of the book."

On screen time, the author argues the concern is not AI-specific. Schools are already reversing course on screens generally. Maine's laptop initiative went fifteen years without moving test scores. A Kansas middle school quietly took its Chromebooks back. Sweden is spending a small fortune replacing tablets with textbooks. That correction started before generative AI showed up. The author does not argue for more screen time; the screen time already exists, mandated by one-to-one programs schools are now walking back. "The real question was never whether a kid is in front of a screen, they already are. It's what they're doing there. AI doesn't add a screen to anyone's day. Used well, it's one of the few things that can make the screen time already sitting in a kid's backpack actually worth something."

The author concedes the human connection concern: more AI in a classroom could mean less human connection, and connection is a big part of what makes school work at all. But the context is important: only 22% of secondary students say most or all of their teachers make an effort to understand their life outside of school, an all-time low down from 43% right after the pandemic hit. That decline happened between 2020 and 2022, years before generative AI showed up. "The relationship problem already existed. Blaming AI for a shortage that predates it lets the actual cause off the hook." The author notes that Dan Meyer's observation—that Khanmigo doesn't love a kid, it can't—is true and worth taking seriously. But "the honest version of that argument cuts the other way too. Every hour AI gives a teacher back from grading and paperwork is an hour that can go toward the one thing AI genuinely can't do: sitting with a kid and paying attention." Whether that hour actually gets spent on connection is a choice, not a guarantee, but it is a choice most teachers do not currently have the time to make at all. In-school connectedness has measurable effects on resilience, stress, and academic outcomes, and the author pledges: "as this technology reshapes the classroom, it's one thread we can't afford to lose."

On equity, the author says the fear is heard constantly from people who consider themselves progressive on education: AI will help kids who already have advantages and leave everyone else further behind. The author respects the instinct but disagrees, at least partially. There is a real academic debate. Plenty of serious people worry that AI adoption in well-resourced schools is widening the gap for rural and high-poverty districts that lack devices or connectivity. That specific problem—raw access—is completely legitimate, the author says, "more than cheating, more than privacy." Implementation is not free; adaptive, personalized systems can run schools tens of thousands of dollars before training staff and maintaining the system, and pricing is still moving fast and not always down. If the gap between what a wealthy district can afford and what a poor one can afford decides which kids get the benefit, "that's a real failure. It's a policy and funding problem, and one worth watching closely, not waving away."

But the argument the author actually hears most often is not about access. It is the claim that using AI at all is somehow unfair because it helps kids who already have advantages. "Even if that premise were true, I'd still reject it: 'I'd rather nobody get the benefit than let some kids get it before everybody can' is a great way to guarantee a permanent race to the bottom." But the author thinks the premise itself gets it backward. "AI is one of the only things I've seen that lets a kid without a lot of resources, or without a grownup who can help polish an essay, or without any perceived natural talent for drawing, produce something they're genuinely proud of. It's a bridge, not a moat." The author recently met with Brazil's Secretary of Education, and one of the points made was that part of the Ministry's interest in AI is in how it can function as a democratizer: a way to narrow the equity gap, not widen it. "It was thrilling to hear a government official responsible for tens of millions of students recognize and articulate the opportunity."

Finally, on privacy, the author acknowledges this fear deserves the most nuance and is real. Kids' data ending up in the wrong places, FERPA and COPPA straining against how these tools collect and process information, and a startling number of teachers—under half—having received any training on using AI safely with student data, all represent genuine gaps. But the author will not accept treating this as an AI-specific problem, because it is not. "Every school already handed a mountain of student data to Google Classroom, to whatever LMS the district picked, to every ed-tech vendor with a signed data-sharing agreement, and largely didn't ask hard enough questions about any of it. If we're suddenly worried about where student data goes, good. We should have been worried the whole time."

The author's central claim is that panic over AI in education conflates two separate issues: the real harms of AI (data privacy, implementation cost disparities) and pre-existing educational failures that AI has simply made visible. The "cheating crisis" described by faculty surveys (78% report increased cheating, 73% have handled AI-related integrity cases) is real, but the proposed solution—AI detection software—is worse than the problem it claims to solve because it systematically misfires against non-native English speakers. This points to a broader theme: AI is not the cause of breakdown in education; it is the mirror held up to one.

The author also addresses the "critical thinking panic"—the fear that students will outsource their thinking to AI—by drawing a parallel to the calculator panic of the 1970s. Calculators did not destroy mathematical reasoning; they shifted where thinking was spent. Similarly, AI does not eliminate critical thinking; it redefines what critical thinking looks like. Evaluating an AI's output, iterating prompts, and knowing when to trust or reject the result are themselves critical thinking exercises. The author's "show your work" framing is the key: when the product alone no longer proves learning, the process becomes the measure.

On fears about screen time, human connection, and equity, the author argues these problems predate generative AI and are not made worse by it—and in some cases, may be improved. Student-teacher connection declined sharply between 2020 and 2022, years before generative AI entered classrooms. Schools were already removing Chromebooks before AI showed up. The equity gap is real, but it is a funding and policy problem, not an AI problem; and the author notes that Brazil's education ministry views AI as a potential democratizer, a tool to narrow gaps rather than widen them. The author's framing is pragmatic: the real threat is not AI itself but the policy failure to ensure equitable access to it.

No comments yet. Be the first to share your thoughts!