Op-ed

Personalized learning needs to go

I could argue that personalized learning should disappear because companies are trying to replace teachers—their empathy, judgement, and mentorship—with algorithms and screens. I could warn about the dangers of privatizing education and handing sensitive learning data to closed corporations. Those are important points, but there is a simpler, deeper reason to reject the current obsession with personalized learning: it undermines the social nature of how humans learn and develop.

As an evolutionary biologist, I study why people behave the way they do. Humans are inherently social creatures. Our evolutionary past and modern behaviour both show we crave connection. That craving helps explain the runaway popularity of social apps and games: we seek interaction, recognition, and collaboration. Personalized learning as it is often implemented runs counter to that instinct. It encourages isolation at a formative stage of life, replacing shared inquiry and face-to-face exchanges with individualized interfaces and solitary tasks.

At first glance, the promise of personalized learning is seductive. An adaptive system could keep each student in the so-called “flow” state—never bored, never overwhelmed—by delivering precisely the right task at the right time. But learning alone is hard. Without intrinsic motivation and experienced guidance, solitary practice becomes daunting and can lead to disengagement and even depression. Social interaction is not an optional extra; it shapes motivation, identity, and the skills we’ll need in the workplace. There is no future job where people always work alone—collaboration, empathy, and communication are essential.

There are two further, practical problems with current models of algorithmic personalization. First, machine learning models are only as good as the data used to train them. Early datasets often come from a narrow slice of the population—affluent, well-resourced schools—because those institutions can afford to pilot new technologies. That skews the algorithm and risks producing recommendations that do not work for diverse students. In other words, biased input yields biased outputs.

Second, an individual cannot be reduced to a pattern of responses. Emotions, context, and lived experience profoundly influence how students perform and engage. A single test score or click pattern does not capture a child’s curiosity, fear, or need for support. Teachers, through empathy and experience, recognise when a student needs encouragement, conversation, or even a hug rather than another worksheet. That human sensitivity is not something an algorithm can reliably replicate.

This is not an argument to reject technology wholesale. On the contrary: technology can be powerful when used to amplify human teaching and to make scientific inquiry accessible, social, and joyful. Instead of isolating students, technology should invite them to explore together, take risks, and practice the messy work of doing science—forming questions, collecting data, and interpreting results.

Natural-born scientists

Children are born to experiment. They learn by trying, failing, and trying again. From their earliest days they observe, manipulate, and test hypotheses about the world. Scientific thinking is essentially curiosity shaped into inquiry: what happens if I push this? Why did that change? Yet many school experiences put theory and nomenclature before practice, creating a barrier that intimidates students before they fall in love with discovery.

True scientific literacy should begin with doing. Many scientific phenomena are invisible or abstract: you don’t see evolution happen, you infer it from trait changes over time; you don’t see molecular processes, you measure their effects. Learning how to decide what to measure, how to collect reliable data, and how to visualize results comes through repeated practice—not only lectures.

But running experiments can be logistically hard in classrooms: equipment needs setup, students need training, and teachers must manage complex procedures. That is where thoughtful technology can help by simplifying the tedious parts—data capture, visualization, and basic analysis—so students can focus on asking questions and interpreting findings.

Imagine a classroom where play and exploration come first, where students use familiar mobile devices to gather and share data, and where teachers spend their time guiding discussion rather than managing apparatus. Would that reduce fear of science? Would it spark curiosity and make students comfortable debating results and proposing new tests? I believe it would.

Designing technology that fosters social science learning

Research shows students feel comfortable with the mobile devices they use daily. Rather than fighting that familiarity, we should harness it to support collaborative inquiry. At arludo, our approach uses mobile apps to get students working together and thinking like scientists. Because students already know how to use their devices, lessons begin quickly and naturally.

As learners interact with apps and with each other, their choices, successes, and failures are collected and anonymously visualized in real time for the class. In doing so, students become scientists without even knowing it: they design experiments, collect data, and see immediate visual feedback. We make the hardest parts of science—experimental design, data collection, and visualization—intuitive, social, and fun.

That shift lets teachers lead reflection and discussion. Instead of setting up and policing activities, teachers ask questions: What does this chart show? Why is there variation? What hypotheses might explain the pattern? How could we test those ideas? Because the data are generated by the students themselves, discussion is personal, relevant, and engaging.

We have developed over twenty mobile applications that cover a range of scientific concepts. Students can run augmented-reality experiments with virtual crabs to learn about animal contests and experimental design, role-play birds to study mate attraction, simulate predator–prey interactions to understand evolutionary pressures, or act as pirates to explore the tragedy of the commons. Each app presents a concept in a memorable, interactive way and invites social reflection.

We also collaborate with researchers to align activities with real scientific questions, giving students a meaningful connection to contemporary research. In short, we want to free teachers to be the empathetic leaders of inquiry while using technology to remove technical barriers to exploration.

Michael Kasumovic is an Associate Professor of Evolutionary Biology at UNSW Sydney, where he researches how social environments and interactions influence self-perception and behaviour. He is also the founder of arludo, a company dedicated to increasing interest in STEM and improving scientific thinking for students of all genders, ages, and socioeconomic backgrounds.