The Gravity of the Present
On the Problems of Imagining the Future of Higher Education
Every serious conversation about the future of higher education begins with a quiet act of sabotage. We picture what the university could become, and almost immediately, the present pulls our imagination back to earth. We envisage better lectures instead of asking whether the lecture should survive at all. I think it should, by the way, but that’s not the point: the present has gravity, and like all gravity it works invisibly - bending the trajectory of every idea back toward what already exists.
This is why most visions for the future are really just tidied versions of the present. They are designed forward: take what we have, extrapolate, move, optimize, and call the result transformation. But forecasting from the present can only ever produce a refined present. The genuinely new has to begin with the future you actually want - for instance, the kind of higher ed institution students and alumni would find worth belonging to - and then you reason backward to what should happen today and in the years ahead for that future to arrive.
Generative AI is making this kind of thinking harder than ever before. With each new model launch, we’re told that it is the most powerful instrument for designing forward ever built. It has arrived at precisely the moment higher education most needs to design backward. Trained on what has already been written and done, generative AI will answer almost any question about the future with a fluent, confident composite of the present. The objection almost writes itself: surely the machine can backcast too, if you only prompt it well. And it probably can, in the mechanical sense. But its raw material is the past, and its instinct, if a statistical inclination can be called that, is regression toward what is already typical.
If we ask a generative AI to imagine the university of 2040 it will produce the university of 2025 with optimized production values. The tool does not prevent backward thinking; it simply makes forward thinking effortless. The discipline this moment demands is to decline the effortless motion the machine is so good at, and to attempt the slow backward act it is worst at: beginning from a future that exists nowhere in what it was trained on.
On Futures and Risk
A well-imagined future is not the same as a future that happens. Clarity is not motion, and change rarely happens because someone explained the future well. People, and perhaps especially academics, are simply not persuaded to move by an elegant roadmap, but rather when they can see themselves inside the future being described, and when they like what they see. Therefore, the work is not to explain the destination more persuasively. It is to make the destination right for the people who will inhabit it.
Students did not begin using generative AI because a committee approved it; they adopted it because, almost overnight, it was easier, faster, and seemingly smarter - short term, at least - than the old way of doing the work. This is gravity inverted, an example of how removing friction leads somewhere new, and the reflexive response has been to push back against it with bans, detection software, and narratives of morality and ethics.
Higher education typically concentrates on delivering consistent academic outcomes (research and graduates), minimizing deviations, and streamlining operations. This focus is only natural: in most institutions, it is what leaders and academics are encouraged and rewarded to do. With exceptionally long product lifecycles in higher education (think about how long it takes to produce a graduate student, let alone an academic), it is the prevailing mode of thinking.
The money flowing in from outside pulls the same way, structurally. External research funding increasingly rewards the foreseeable: projects with defined deliverables, predictable milestones, and impact that can be promised in advance. The proposals most likely to be funded are, almost by design, the ones least likely to surprise anyone - so ambition gets trimmed to fit the grant, and the genuinely uncertain, high-risk idea, whose payoff cannot be specified up front, learns not to apply. Caution inside the institution and caution in the funding system reinforce each other until risk aversion stops looking like a strategic choice and starts looking like the default.
This type of consistency is itself a form of gravity; it keeps everyone anchored to the safety of the present until the danger feels fully retired, which is never.
Why We Keep Pushing
Pull wins over push any day of the week. Imagined forward, generative AI becomes a better plagiarism detector: the existing assignment, the existing credential, the existing lecture, all defended against a new threat. Imagined backward, it forces the question the take-home essay was always quietly avoiding - is this work really assessing what we think it is?
That same pull toward the present is visible in the research literature, too. Surveying the recent literature on assessment in the age of generative AI, Phillip Dawson points out that the overwhelming majority frames it as a matter of cheating and integrity, while only a handful raise the question of validity at all. A critical review he cites, by Ramiz Ali and Jerry Maroulis, sorts the field into three stances: risk-oriented, rule-based, and design-focused. The first two are push, defending the old assessment by policing the new threat; only the third asks the backward question and redesigns the task so the threat has nothing to grip. This uneven balance is not a coincidence; it is gravity, pulling even the specialists back toward guarding what already exists.
The gravity never switches off, and the work of imagining the future of higher education is, in the end, a daily discipline of refusing to let what exists define what could be. A future built by motion has to reward motion. Credit the experiment that pointed somewhere true even if it failed; promote the people who tried, not only those who arrived. Movement compounds, perfection paralyzes.
The goal is not to be right all at once but to make the next right step easier than the last.
Easier said than done, obviously, but there it is.
Fuel the next post ☕️ - buy Jeppe a coffee
