AI Won’t Save Education If Students Aren’t Motivated

The real challenge in modern education isn't artificial intelligence. It's human motivation.

Artificial intelligence is reshaping education faster than most institutions can respond. Schools are writing AI policies. Universities are redefting academic integrity. Teachers are rethinking assignments. And in all of this urgency, the conversation keeps circling around one question: what can AI do?

Rarely does anyone stop to ask a more important one: what do students actually need?

This article argues that the more important question is psychological, not technological. And until we take it seriously, we will keep designing AI integration around capability rather than around the learners as it is supposed to serve. Before asking how AI should be used in classrooms, educators need to understand what actually drives learning. The answer, supported by decades of research in educational psychology, is motivation and specifically, the conditions that produce it. Student motivation is not a soft variable sitting beside the “real” work of education. It is the mechanism through which learning happens at all. Get it wrong, and no amount of intelligent technology fills the gap.

The disengagement problem we keep ignoring

Student disengagement is not a new crisis, but it is a deepening one. A 2023 Gallup survey of U.S. students found that only 47% of high school students felt engaged in school — a figure that had been declining steadily for years (Gallup, 2023). Globally, the OECD’s PISA data consistently shows that a significant proportion of students lack intrinsic interest in their schoolwork, completing tasks out of compliance rather than curiosity (OECD, 2023).

Educational psychologists map motivation along a continuum. At one end sits intrinsic motivation — doing something because it genuinely interests you, because it connects to who you are and what you value. At the other end is amotivation — not low motivation, but its complete absence. A student who is amotivated does not see the point of the activity, does not believe their effort will lead anywhere meaningful, and is essentially going through motions (Deci & Ryan, 1985).

In between are various forms of extrinsic motivation: working for grades, for approval, to dodge consequences. These are not useless — external regulation can get things done — but research is consistent that intrinsic motivation produces outcomes that extrinsic motivation simply cannot: deeper conceptual understanding, longer-term retention, greater creativity, and better psychological wellbeing (Ryan & Deci, 2000).

So before asking where AI fits in classrooms, we should ask: where on this continuum does AI use push students?

Self-Determination Theory: what the research actually shows

The most rigorously tested framework for understanding human motivation is Self-Determination Theory (SDT), developed by Edward Deci and Richard Ryan over four decades of cross-cultural research and more than 800 published studies (Ryan & Deci, 2017). It is, by any reasonable measure, the dominant empirical theory of motivation in education.

SDT identifies three basic psychological needs. When classrooms meet them, intrinsic motivation grows. When they frustrate them, amotivation follows.

Autonomy — engaging because you choose to, not because you’re being watched. Autonomy-supportive teaching environments consistently produce higher intrinsic motivation and better academic persistence than controlling ones (Reeve, 2009; Vansteenkiste et al., 2012).

Competence — growing through your own effort. Hattie’s (2009) synthesis of 800+ meta-analyses found that feedback supporting genuine competence is among the most powerful influences on learning outcomes. Competence earned feels fundamentally different — psychologically and cognitively — from performance produced by a tool.

Relatedness — feeling seen and connected to teachers, peers, and the subject itself. Osterman (2000) found that students’ sense of belonging robustly predicts motivation, engagement, and achievement. Learning happens within relationships. It is shaped by them.

These aren’t aspirational ideals. They are baseline psychological requirements — as foundational to learning as sleep is to cognition. Remove them, and motivation deteriorates regardless of what tools surround the student.

What AI does to each of these needs

When AI functions primarily as a task-completion tool — generating essays, solving problem sets, producing work students submit as their own — it does not merely raise ethical questions. It structurally undermines all three conditions that motivation depends on.

Autonomy erodes. Prompting AI to write your essay means directing a machine, not making intellectual choices. Deci et al. (1999) found across 128 experiments that externally controlled task completion consistently reduces intrinsic motivation.

Competence is bypassed. [Bjork and Bjork (2011)](http://Bjork, E. L., & Bjork, R. A. (2011). Making things hard on yourself, but in a good way: Creating desirable difficulties to enhance learning. Psychology and the real world: Essays illustrating fundamental contributions to society, 2(59-68), 56-64.) call productive struggle “desirable difficulty” — it’s uncomfortable, but it’s the mechanism through which durable learning is built. AI that removes the struggle removes the learning.

Relatedness disappears. There’s no negotiating with another person’s thinking, no real audience, no shared intellectual space. Fredricks et al. (2004) found social and emotional connection is a core component of sustained engagement — not a bonus.

This isn’t a morality argument about cheating. AI used as a shortcut structurally disrupts the psychological conditions that make learning possible.

Scaffolding versus substitution:

The above critique does not mean AI has no legitimate place in education. The research on learning supports a precise distinction between two fundamentally different uses.

Vygotsky’s (1978) concept of scaffolding describes support that enables a learner to do something they could not do alone — with the explicit purpose of building toward independent capacity. A scaffold is temporary by design. It exists to become unnecessary. AI used as a scaffold might help a student organize their thinking before drafting, access background knowledge to understand a concept, or receive targeted feedback on work they have already produced. Used this way, AI genuinely supports the development of competence.

Substitution is different in kind, not just degree. It replaces the learner’s cognitive activity with the tool’s output. The student does not grow because the growth opportunity was eliminated. Winne (2017) frames this as “offloading cognition” — and the research on repeated cognitive offloading suggests that when we habitually delegate thinking to external tools, our independent capacity in that domain does not stay the same. It diminishes.

Kirschner and De Bruyckere (2017) extend this point by dismantling the myth that students who are digitally fluent are necessarily learning from digital tasks. Comfort with a tool is not the same as learning through it. The two can come apart completely.

What this means for teachers

Educational psychology does not leave teachers without guidance here. It offers quite specific direction.

Reeve (2009) identifies autonomy-supportive teaching behaviors that consistently predict higher student motivation: providing genuine rationale for tasks, taking students’ perspectives seriously, offering real choices within appropriate structures, and using language that informs rather than controls. None of these behaviors require avoiding AI. But they do require that teachers think carefully about whether AI use, in any given context, supports or undermines students’ experience of volition.

Hattie and Timperley (2007) show that effective feedback — the kind that builds competence — targets the process of learning, not just its outcomes. A teacher who designs AI use around how students think, rather than around what the AI produces, is far more likely to preserve the conditions under which competence actually develops.

In my own research, which examined teacher motivation during project-based learning in AI-integrated settings, the most motivating environments shared a consistent feature: students encountered genuine problems where their own effort visibly mattered. AI was sometimes present as a resource. But it was not positioned as the answer. The psychological conditions for motivation — autonomy, competence, relatedness — were actively protected, not left to chance.

The question that needs to come first

The mainstream conversation about AI in education is dominated by questions of access and capability. Can all students reach these tools? What can the tools do? These are legitimate questions. But they are downstream of a more foundational one: what does a student need in order to actually learn?

The research answers this question with unusual consistency. Students need to experience their engagement as self-chosen. They need to feel that their effort produces real growth. And they need to feel that their learning connects them to people and ideas they care about. These are the conditions that produce intrinsic motivation. And intrinsic motivation is not just a pleasant feature of good classrooms — it is the mechanism through which deep, durable learning occurs (Ryan & Deci, 2000).

AI does not generate these conditions. In many of its current uses in classrooms, it quietly undermines them. That does not make AI a threat to education. It makes it a tool that requires psychological literacy to use responsibly.

The educators who will genuinely serve students in the years ahead are not necessarily those who adopt AI fastest or resist it longest. They are those who understand — with real clarity, grounded in research — what motivation is, why it matters, and what it takes to protect it when powerful shortcuts are available.

That understanding does not start with technology. It starts with the human being sitting in front of it.