Why look back
Predictions about online learning have been made since the web became widely used, and enough time has passed to compare some of them with what happened. This article reviews common forecasts from roughly the 1990s onward, sorts them into those that largely came true, those that came true in part, and those that did not, and draws lessons. It stays at the level of widely observed patterns. Specific dates, figures and quotes are avoided where they cannot be checked here.
Predictions that largely held up
Online delivery would become ordinary. Early commentators expected web-based courses to spread across universities, colleges and workplaces. They did. Learning management systems, recorded lectures and online submission are now common across education, though the extent differs by institution.
Learning would reach people who had been excluded by distance or schedule. Distance education existed long before the web, but online delivery broadened it. Working adults, carers and people far from campuses have been among those using it.
Content would become cheaper to distribute, and some would be free. The expectation that educational material could be shared at close to zero marginal cost proved sound. Open licensing and free course materials followed, and freely available explanations and tutorials are everywhere. What the forecasts underestimated was how much quality and support still cost.
Employers and learners would turn to short, skills-focused study. This has been borne out in the growth of short courses and certificates, although how much employers value them varies.
Predictions that came true only in part
“Free courses will open higher education to everyone.” Free online courses attracted very large numbers of sign-ups. But completion rates were commonly low, and early participants were often already educated. The forecast was right about reach and optimistic about who benefited. Support, motivation and recognition turned out to matter as much as access.
“Personalized learning will tailor study to every student.” Adaptive practice tools exist and have some research support in structured subjects. The strong version of the forecast, that software would understand each learner and adjust everything, has not been shown to hold generally. Personalization tends to be narrower than promised.
“Online learning will cut costs dramatically.” It can lower some costs, such as textbooks and travel. But building and supporting good courses is expensive, and many online programs have not been much cheaper for the institutions running them or for students. Savings, where they occur, depend on scale and design.
“Remote learning will become the default after 2020.” The shift to remote teaching in 2020 led some to expect a permanent move. In practice, many places returned toward in-person teaching while keeping some online tools. Hybrid practice has stayed, but a total change did not.
Predictions that did not materialize
Virtual universities would quickly outcompete traditional ones. Around the turn of the century, several well-funded ventures were launched on the expectation of rapid growth, and a number of them closed or were scaled back. Established institutions with reputations, campuses and credentials proved resilient and moved into online provision themselves.
Institutions would collapse. Forecasts that a large fraction of colleges would vanish within a few decades have appeared regularly. Some institutions have struggled for financial and demographic reasons, but the sector has not been displaced by online competitors.
Technology would replace teachers. Each new medium prompted this claim. Support roles changed, but teaching and tutoring remain central, and studies of online learning often find that human contact is a key factor in persistence.
Patterns in the record
- Reach was predicted better than impact. Forecasts about how many people could access learning fared better than forecasts about how much they would learn.
- Direction beat timing. Saying that a change would occur was often right; saying when, and how fast, usually was not.
- Replacement claims fared worst. Statements that something would be swept away rarely came true, while claims that new options would grow alongside old ones did better.
- People and institutions were underestimated. Motivation, trust and habit slowed some changes. Established institutions adapted rather than vanished.
- Money was underestimated. Producing quality learning at scale, and sustaining it, took more resources than many models assumed.
Lessons for reading new forecasts
- Prefer modest claims about direction over precise ones about timing.
- Distinguish access from outcomes. A tool that reaches many people has not yet shown that they learned.
- Be skeptical of replacement narratives and look for how existing institutions might adapt.
- Ask who pays and who supports learners, since these questions decided many earlier cases.
- Check whether similar forecasts were made before, and what happened.
- Expect surprises. Some of the most important developments, such as mobile-based study, were not the focus of early predictions.
A caution about hindsight
Looking back is useful but easy to do unfairly. Past forecasters worked with less information, and outcomes depend on events nobody could foresee, such as economic shocks or health emergencies. A forecast can be reasonable and still fail, or be careless and still succeed by luck. The point of the record is to improve our questions, not to ridicule earlier writers.
New forecasts about online learning will keep coming. Applying the lessons above, favoring cautious, evidence-linked claims and noticing hype, gives a fair chance of reading them sensibly, though the future will remain uncertain.