The AI Moment

The Evolution of EdTech — Part 9 of 13

By Clarence Stephen | April 2026

2,300 words | 9 min read

At 1:00 PM Eastern on November 30, 2022, OpenAI quietly released a chatbot called ChatGPT. No press conference, no Super Bowl ad, no keynote. Just a blog post and a URL.

Within five days, a million people had signed up. Within two months, 100 million. Fastest-growing consumer application in history. And nowhere was the impact felt more immediately than in education.

Within weeks, students discovered that ChatGPT could write essays. Not passable essays — competent ones. Essays that marshaled evidence, constructed arguments, and produced prose that was, if not elegant, solidly workmanlike. B+ in most American high schools. A respectable B in many college courses.

If you've been reading this series, you already know what happened next: the educational establishment panicked.

The Great Ban

The New York City Department of Education — 1.1 million students, the largest district in the country — banned ChatGPT from school networks and devices in January 2023. Six weeks after launch. Los Angeles Unified followed. Seattle banned not just ChatGPT but six additional AI writing tools.

The logic was straightforward: if students can use a machine to write essays, then essay assignments are meaningless. If essay assignments are meaningless, a huge portion of how we assess learning is broken. So ban the machine.

Logical. Also futile.

Students accessed ChatGPT on personal phones. Used it at home. Used VPNs to bypass school networks. Surveys found roughly one in four teachers had caught students using AI on assignments. The actual rate was certainly far higher — those were the students who got caught.

We Have Been Here Before

The AI panic should feel familiar by now.

Teaching machines in the 1920s: critics feared mechanized learning. Television: fear of passivity. Personal computers: fear of teacher replacement. The internet: fear of unreliable information. Smartphones: fear of distraction.

Every time, the same sequence: fear, banning, grudging acceptance, integration. Every time, the technology turned out to be neither as transformative as boosters promised nor as destructive as critics feared.

But here's where ChatGPT broke the pattern. Previous technologies threatened to change how students access information or receive instruction. ChatGPT threatened to change what it means to demonstrate knowledge. It struck directly at how education assesses learning.

Think about that. If a student can produce a well-reasoned essay without actually reasoning — if the output of learning can be generated without the process — then the entire assessment system built on written work is unreliable. This wasn't a technology challenge. It was an epistemological crisis.

The Plagiarism Wars

What followed was predictable: a frantic search for detection tools.

Turnitin rushed to add AI detection. GPTZero, built by a Princeton student named Edward Tian, launched in January 2023 and quickly attracted millions of users. OpenAI itself released an AI text classifier — then quietly retracted it.

None of them worked reliably.

The problem was structural. AI-generated text isn't like copied text. Plagiarism detection works by comparing submissions against a database and flagging matches. AI-generated text is original — it's never existed before. And in many cases, it's statistically indistinguishable from competent human writing.

Detection tools produced unacceptable rates of false positives (flagging human writing as AI) and false negatives (missing AI text). The false positives were devastating: students who'd written their own work got accused of cheating. Those accusations fell disproportionately on non-native English speakers, whose writing patterns triggered detection algorithms — adding an equity dimension to an already messy situation.

By mid-2023, the detection arms race was clearly a losing proposition. Here's the thing: the assessment problem couldn't be solved by better detection. It required better assessment.

The Teachers Who Said Yes

While the ban debate consumed institutional oxygen, something more interesting was happening in actual classrooms. Thousands of teachers — often without institutional support, sometimes in direct defiance of district policy — started experimenting with AI as a teaching tool.

They saw something the ban-first crowd missed: ChatGPT wasn't just a cheating tool. It was a thinking tool.

A high school English teacher in Texas used ChatGPT to generate first drafts, then had students critique and improve them — teaching analysis and editing more effectively than traditional red-penning. A middle school science teacher asked students to fact-check ChatGPT's explanation of photosynthesis. The AI made errors students could identify and correct — a better scientific reasoning exercise than any textbook problem set.

A special education teacher generated reading passages at different difficulty levels — a task that used to take hours. An ESL teacher created conversation scenarios tailored to individual proficiency. A math teacher asked students to write prompts that would make ChatGPT solve problems correctly, and discovered that composing effective prompts required deeper mathematical understanding than solving the problems directly.

These teachers weren't naive about the risks. They worried about cheating, over-reliance, content quality. But they understood that AI wasn't going away, and that preparing students for a world with AI meant teaching them to use it thoughtfully — not pretending it didn't exist.

The Deeper Question

Beneath the plagiarism wars, AI surfaced a question education has been dodging for decades: what are we actually assessing, and why?

The five-paragraph essay — backbone of American writing instruction since roughly the 1960s — was designed to test whether a student could organize thoughts, construct arguments, and communicate clearly. It was a reasonable proxy for critical thinking in an era when the only way to produce such an essay was to actually think critically.

AI broke the proxy. A student could now produce the output (a polished essay) without the process (critical thinking). But this exposed something important: the output was never the point. The process was always the point. The essay was just the closest available window into a student's mind.

The teachers who adapted fastest shifted focus from output to process. Instead of "did you write this?" they asked "how did you arrive at this argument?" Instead of grading a finished product, they assessed the journey — drafts, revisions, reasoning, dialogue.

Look. This is better pedagogy regardless of AI. The five-paragraph essay was never a great assessment of critical thinking. It was the best tool given the constraints. AI broke the old tool and forced a conversation about building better ones. Some crises are productive.

The Corporate Response

The EdTech industry responded with the velocity and restraint of a gold rush.

Within months, hundreds of "AI-powered" educational tools appeared. Some were genuinely innovative. Many were ChatGPT with an education-themed wrapper and a higher price. "AI-powered" became the EdTech equivalent of "organic" — applied generously, defined loosely, used primarily as a marketing badge.

The serious players moved fast. Khan Academy announced Khanmigo, an AI tutoring assistant built on GPT-4, in March 2023. Duolingo integrated AI conversations. Google and Microsoft added AI features to their education platforms. MagicSchool AI launched practical tools for teachers — lesson planning, rubric generation, text leveling — addressing real needs rather than hypothetical ones.

The gold rush raised familiar questions. Would AI tools widen or narrow the equity gap? Help teachers or replace them? Improve learning or just improve the appearance of learning?

The answers depend, as they always have, not on the technology but on implementation. By whom, for whom, and with what support.

Where We Stood

By end of 2023, the landscape had shifted. New York City reversed its ban, acknowledging that blocking the technology was neither effective nor desirable. Other districts followed. The conversation moved from "should we allow AI?" to "how should we use AI?"

But the core tensions remained unresolved. The assessment crisis was real. The equity concerns were real. The potential for both transformation and harm was real. And the pace of change was accelerating in ways that made thoughtful, evidence-based policy development feel almost quaint.

The AI moment in education wasn't a single event. It was the beginning of a phase change — a fundamental shift in the relationship between technology and learning that will play out over decades. The teaching machine era lasted a generation. The internet era lasted a generation. The AI era moves at the speed of software updates.

What happened next — AI tutoring systems backed by actual evidence — is the subject of the next chapter. And for the first time in this series, it's genuinely cause for optimism.

Next in the series: "Adaptive Learning Gets Real" — when AI tutoring stopped being a promise and started being a product.

Sources

Clarence Stephen is the founder of Readify, an AI-powered literary platform. Yale B.Sc. Physics. Previously: Tiger Global, Morgan Stanley. Learn more at ireadifybooks.com

What could AI do for your classroom?

Readify uses local AI to adapt reading material to each student's level — generating illustrations, simplifying vocabulary, and tracking comprehension in real time. No student data leaves the building.

Bring Readify to Your School Try the Platform

Liked this article?

Get the full Evolution of EdTech series delivered to your inbox, plus weekly insights on AI in education.

Subscribe to Readify Branches