When ChatGPT dropped in November 2022, higher education administrators did what they always do: convened committees, drafted policies, and waited to see which way the wind would blow. But a new analysis from academic and developer Nicolas Saint-Lager argues that treating AI as "just another MOOC" might be exactly the wrong move—and this time, the usual playbook won't cut it.

The Three Tech Shocks That (Mostly) Whiffed

Saint-Lager breaks down how higher ed absorbed three major disruptions since 2000: the web and internet explosion, big data and data science hype, and MOOCs promising to democratize education. Coursera, edX, Udemy—these platforms still exist, sure. But universities are doing just fine. The author's firsthand experience at emlyon business school under Bernard Belletante's leadership is telling: even visionary adaptation meant adding new programs (data science tracks, Makers Labs) rather than reimagining the core lecture-and-seminar model that's persisted for centuries. The web transformed operations—email, learning management systems, video conferencing—but the classroom itself barely flinched. Learning analytics and adaptive tools exist, but group-based instruction remains the foundation. Saint-Lager's conclusion is blunt: these previous shocks were absorbed as topics to dissect in class rather than forces reshaping how that class functions.

Why AI Breaks the Pattern

Here's where it gets interesting from a hacker-culture perspective. The author argues AI isn't just another layer on top of existing infrastructure—it's fundamentally different because it expands creativity itself, not just the space for creative expression. When an LLM can generate in minutes what a skilled student produces in hours or days, you can't just add "AI Ethics 101" to the curriculum and call it handled. The four dimensions that set AI apart: it's profound (exceeding human baselines on cognitive benchmarks), broad (LLMs emulate virtually any domain we request), systemic (environmental strain from data centers, geopolitical concentration in US/China hands, potential workforce displacement), and rapid. ChatGPT hit 1 billion monthly active users in roughly three years—compare that to the two-to-three decades it took for web adoption to become widespread.

The Geopolitical Dimension Nobody's Talking About

This is where Saint-Lager's analysis gets genuinely unsettling from an infrastructure standpoint. While the early web remained relatively open across borders, frontier AI services are concentrated in two regions: the United States and China. European schools face genuine dependency risks. When Anthropic rapidly released and shut down Fable 5 in June 2026 following US government directives restricting non-citizen access, it exposed how fragile "open" AI access actually is. Mistral AI and Black Forest Labs offer credible European alternatives in language models and visual intelligence respectively—but they can't match the frontier capabilities coming from San Francisco and Beijing. For institutions that assumed cloud-based services meant vendor independence, this should be a wake-up call.

The Existential Question Hanging Over Campuses

Beyond operational concerns, Saint-Lager doesn't flinch from bigger questions: AI increases potential for harm in cyberattacks and bioweapon design, while figures like Geoffrey Hinton have signed statements flagging plausible scenarios of loss of control or existential risk. This isn't science fiction from a 1990s hacker novel—it's being debated seriously in academic circles now.

The Bottom Line

Universities that treat AI policy as checkbox compliance work are playing checkers while the board's already evolved into something unrecognizable. Previous tech shocks tested higher ed's resilience and found it surprisingly robust—but those shocks never questioned whether learning the skill itself matters when AI can replicate it instantly. Schools need to stop asking "how do we use AI?" and start asking "what does education mean when anyone can generate the output without acquiring the process?" That's a harder conversation, but pretending otherwise is just postponing the inevitable.

Key Takeaways

  • Previous tech shocks (web, big data, MOOCs) transformed operations but left core pedagogy surprisingly intact
  • AI differs fundamentally: it expands creativity itself, not just access to creative tools
  • ChatGPT reached 1 billion MAUs in ~3 years vs. 2-3 decades for web adoption—speed matters
  • Geopolitical concentration (US/China) creates dependency risks for European institutions
  • The existential risk questions AI raises can't be addressed by adding another elective