This article first appeared on University World News on 21 January 2026.
If the university had a ‘Like’ button, it would be a small confession of defeat. A button is a decision protocol: click, register, rank. Platforms are built to harvest these clicks and, in return, arrange the world so that more or less of the same will come. Their basic code is attention/not-attention.
Universities are built to harvest something else: the slow conversion of confusion into capability. Their code is truth/not-truth in the domain of knowledge, and qualification/non-qualification in the domain of education. These codes can borrow from each other, occasionally flirt, sometimes quarrel – but when attention becomes the chief arbiter of academic value, the court changes jurisdiction. You get verdicts efficiently, but on the wrong case.
We may as well be honest. The platform has already strolled into the campus wearing a lanyard: the learning management system with its heat maps of ‘engagement’, the dashboards that colour courses in green or red, the course evaluation stars that twinkle at budget meetings, the video player that counts watch-time as if comprehension were a cousin of endurance. All of it feels empirical, humane, modern. And it is. It also quietly inverts who decides what counts as learning.
Platforms thrive by provoking short cycles of response. Universities thrive by cultivating long cycles of reflection. This is not a moral difference but a temporal one. The semester is slow because thinking, like fermentation, resists haste. The feed is fast because attention, like sugar, is cheap and addictive. Mix the two, and you breed an organism that metabolises sugar while declaring loyalty to nutrients. It will smile while it starves.
Consider the innocuous phrase now found in almost every internal memo: engagement. Engagement is to learning what mood is to romance: relevant, not dispositive. Painfully, learning often requires strategic disappointment.
A well-designed course introduces selective difficulty – concepts that resist, problems that refuse to be completed in one sitting, texts that do not repay the first reading. This difficulty is not gatekeeping; it is the ladder that later becomes muscle memory. When engagement metrics govern course design, difficulty is trimmed where it offends. The ladder is lowered to the point where it no longer lifts.
The consequences are systematic. First, the curriculum smooths toward the trend – modules on ‘AI for Everyone’ multiply while logic withers, because one yields clicks with novelty and the other with obligation.
Second, assessment drifts from provenance to product – grading what is turned in, not how it was produced – since provenance is hard to quantify and dashboards prefer numerators.
Third, academic time is colonised by micro-rewards and perpetual notification: the course becomes a sequence of nudges that maximises return visits instead of a sequence of experiences that maximises transformation.
Fourth, faculty become content creators in their own house, anxiously refreshing the watch-time of lectures as if boredom were a legal violation.
None of this is a call to smash the servers. The tools are not the problem; their governance is. Plumbing is wonderful; we just do not want it running the city council. Digital platforms carry messages and store work; they should not decide what a good message is or what good work looks like. The difference is subtle in code and obvious in outcome: does the platform report to the curriculum, or does the curriculum adapt to the platform’s reporting features?
Making good design choices
To preserve universities as places where truth and qualification remain decisive, we need a few firebreaks – design choices that separate platform logic from academic judgment while keeping the good plumbing.
• Establish a governance firewall for analytics. Analytics are operational tools, not evaluative authorities. Use them to spot who has not logged in, not to conclude who has not learned. Prohibit the use of watch-time, click counts and response latency as proxies for course quality or faculty performance. Yes, build dashboards; no, do not let them drive.
• Reassert curriculum sovereignty. The syllabus is not a menu to be optimised by A/B testing; it is a constitution that distributes difficulty, attention and care. Vendor features cannot mandate pedagogy. Require that substantial shifts in assessment or content go through peer academic bodies, not software default settings. Pilot platforms in time-boxed trials with exit ramps; never wed a tool by policy.
• Disable invisible algorithms by default. Discussion boards should be chronological unless the instructor chooses another order, and the algorithm must reveal its criteria when it does. Turn off autoplay and auto-advance in instructional videos; students are not binge-watching a plot. Replace engagement streaks with reflection milestones. In brief, prefer affordances that respect the student’s agency over those that shape it covertly.
• Make difficulty a declared design element. In every syllabus, include a difficulty map that explains where effort will concentrate and why. Pair hard segments with structured support – office hours, peer sessions, annotated exemplars – so that difficulty is fair, not arbitrary. Evaluate courses partly on whether they deliver this promised difficulty, not merely on whether students liked it. If we must have a star, let it rate the quality of struggle.
• Shift assessment towards provenance and process. Require version histories, oral defences or reflective memos that explain what was attempted, revised and learned. In a world where machines draft, the human skill is to direct, critique and justify. Grade the choreography, not just the dance. This reorients teaching towards judgment – what machines cannot do and students most need to practice.
• Practice index hygiene. Publicly declare which metrics your university refuses to optimise: watch-time, star ratings and response rates among them. Publish what will matter instead: peer observation of teaching, external examiners for capstones and longitudinal performance in later courses. If you cannot say no to a number, the number has already said no to you.
• Add sunset clauses to administrative accretions. Any policy that adds clicks, forms or mandatory modules should expire unless it demonstrably frees faculty time for feedback or student time for study. Otherwise, the organisation that organises itself will keep organising until teaching is the residue that remains after compliance.
Some counterarguments
I anticipate the counterarguments. Without engagement, we lose students. Retention saves budgets. True. But retention is not the same as education, and students remain not only for what entertains but for what changes them. The gym that keeps you attending by hiding the weights is not ethical; it is mercifully short-lived.
The economical choice is to keep students by building capability, which is the only thing that compounds. When graduates can do more than they could before, they will say so – in public, in salary negotiations, in the quiet confidence of solving something difficult. That is the marketing plan that does not decay.
Another worry: students like convenience. Of course. So do we. That is why we should spend convenience on the right things. Use platforms to remove friction in relation to logistics – registration, deadlines, submission – and conserve friction where it produces learning – argument, practice, revision. In this sense, fairness requires difficulty: it gives every student a chance to acquire what cannot be bought as a service.
A third concern: equity. Isn’t difficulty exclusion in disguise? It can be, when difficulty is chaotic, unannounced or defended by mystique. But the remedy is not to remove difficulty; it is to design it. Make the ladder visible. Provide more rungs. Offer scaffolding without lowering the roof. The ultimate inequity is to graduate students into a world that will demand judgment while having trained them to seek a button.
Behind all of this sits a more delicate sociological point. Teaching and learning are forms of double contingency: I expect that you are trying to learn, you expect that I am trying to teach, and we both act under the uncertainty of each other’s next move. Trust bridges this uncertainty.
Platforms promise to replace trust with telemetry. But telemetry cannot tell you whether silence in a seminar is laziness, confusion or careful thinking; it can only tell you it happened. To interpret silence, someone must risk an intervention, perhaps a question, perhaps a pause. This risk is what makes a teacher different from an algorithm. Risk is the price of meaning. What would a university look like that accepts platforms as tools and refuses them as governors?
It would look, I suspect, less like a feed and more like a workshop. There would be fewer notifications and more appointments; fewer badges and more artefacts; fewer leaderboards and more letters of critique. Course evaluations would include narrative feedback read by peers, not just stars displayed for fear.
Faculty would stop apologising for assigning hard texts and start explaining why difficult reading is a form of respect. Students would still watch videos and use AI, but they would be asked, regularly, to defend decisions in their own voice.
Have the courage to disappoint
And administrators? They would hold two ledgers. In one, the numbers that keep the institution solvent. In the other, the commitments that keep it intelligent. Both are necessary; confusing them is fatal. When the numbers threaten to colonise the commitments, leadership would say so, publicly, and accept a little unpopularity; the courage to disappoint is a professional requirement in education, not a character flaw.
If we must add a button for the university, let it say ‘Because’. Not ‘Like’, not ‘Share’, not ‘Subscribe’ – ‘Because’. Because the argument holds. Because the method is sound. Because the difficulty was warranted. Because you changed your mind after considering the evidence.
The ‘Because’ button is slow, hard to press and gives no dopamine. It offers something else: the confidence, years later, to solve what your younger self could not even name.
Newton, I sometimes think, would have made a terrible influencer. Gravity does not trend, and the Principia is a poor candidate for a thumbnail. Fortunately for him – and for us – no one asked him to upload on Thursdays. Let us extend the same courtesy to our students. Keep the tools. Disable the seduction. Build the ladder. And please, leave the ‘Like’ button to the places that sell attention. We have more stubborn goods to trade.