Every few years, some tech evangelist declares that higher education is about to be disrupted into oblivion. The Internet was supposed to make universities obsolete. MOOCs would replace professors. Big data would personalize learning at scale. None of it happened—at least not in the ways promised. Now AI has arrived with the same breathless predictions. But a new analysis suggests this time might actually be different, and not in a good way.

Three Tech Shocks That (Mostly) Whiffed

The blog post by academic and developer Michel Lenoir traces higher ed's recent history of tech disruptions. The Internet and web brought digital content production to everyone—YouTube lectures, shorter attention spans, mobile-first learning habits. Big data promised personalized learning paths through analytics on student data. MOOCs like Coursera, edX, and Udemy offered Ivy League credentials at a fraction of the cost. Each shock generated intense anxiety about university survival.

So Why Are Schools Still Standing?

Here's what's striking: these technologies transformed operations but left the core classroom experience surprisingly intact. Email, learning management systems, video conferencing—useful tools, sure. But the fundamental format of professor-plus-classroom-plus-students remains unchanged after 25 years of 'digital transformation.' Even smartboards get used with moderation, according to Lenoir's observations. The web got absorbed as a topic to discuss rather than a format to adopt.

Four Reasons AI Is Different

Lenoir argues AI differs from previous shocks across four critical dimensions: profound (AI expands creativity itself, not just space for expression—generating in minutes what skilled students create in hours), broad (LLMs can emulate virtually any cognitive task we assign them), systemic (data centers consuming massive electricity and water, geopolitical concentration in US/China hands, potential white-collar unemployment, existential risk discussions from figures like Geoffrey Hinton), and rapid (ChatGPT hit 1 billion monthly active users in roughly three years versus two to three decades for web adoption).

The Geopolitical Wildcard

Perhaps most concerning for European institutions: while the early web remained relatively open across borders, frontier AI is concentrated in just two regions—the United States and China. Anthropic's rapid shutdown of Fable 5 in June 2026 following US government directives restricting non-citizen access illustrates how quickly access to these services can vanish. 'This creates a dependency risk for European schools and universities,' Lenoir notes. Open-weight models exist, but they come with their own significant infrastructure costs.

The Existential Question

Lenoir doesn't shy away from the heavier implications. AI models increase potential harm in cyberattacks and bioweapon design. Some researchers consider AI an existential risk to humanity through plausible scenarios of loss of control or misaligned objectives. 'The web, while not a calm and quiet innovation, certainly never amounted to this level of systemic risk,' he writes.

What Actually Happened With the Last Tech Shocks

For context on why institutions might be tempted to wait this out: emlyon business school under Bernard Belletante's leadership successfully navigated previous shocks by anticipating them—launching data science programs, creating maker spaces, adapting classrooms for hybrid learning. Schools absorbed innovations operationally while their core missions stayed intact. The Gartner Hype Cycle suggests inflated expectations eventually give way to plateaued productivity.

Key Takeaways

  • Previous tech shocks (web, MOOCs, big data) transformed operations but not the fundamental classroom format
  • AI is argued to be more profound, broad, systemic, and rapid than any previous disruption
  • Geopolitical concentration in US/China creates access risks for European institutions
  • ChatGPT reached 1 billion MAU in ~3 years versus decades for web adoption
  • The author's conclusion: we can't safely assume AI will be as 'innocuous' as past tech fads

The Bottom Line

Universities have gotten comfortable assuming the next big thing won't actually change anything. That bet might have worked for MOOCs and learning management systems. But when a technology can generate doctoral-level explanations, create content in minutes that took skilled students hours, and is controlled by two geopolitically rival powers, waiting it out starts to look like rolling dice with loaded weights.