Empowering Self-Regulated Learners in the Age of AI

As a classroom teacher, I reminded my students that the real magic wasn’t in the polished paper but in the messy thinking that produced it. I encouraged “sloppy copies” as the place where breakthroughs happened, celebrated mistakes as opportunities, and told them that scratch paper in math was just as valuable as the final answer. Without the creative friction that comes from wrestling with ideas, learning risks becoming hollow.

The rise and increased use of generative AI threatens to remove that messy thinking from the picture. While AI certainly has its upsides, it is not without fault and it has thrown education into a world of paradoxes. Some educators race to adopt the sharpest detection model to catch misuse, while others encourage students to experiment with AI as a learning aid. At the same time, education tools are racing for ever-more accurate detection systems, even as they embed AI into nearly every corner of the student experience. AI can generate essays in seconds, erasing the need to think. Yet the same technology holds extraordinary potential—if students have the skills to use it wisely. The difference comes down to self-regulation.

Why Self-Regulation Matters More Than Ever

In a landscape where even well-intentioned teaching practices and tools can unintentionally undermine authentic learning, self-regulated learning becomes the thread that holds meaningful engagement together. Self-regulated learning (SRL), as defined by Barry Zimmerman, is the ability for learners to plan, monitor, and reflect on their own learning. Recent research shows that self-regulation is the key factor distinguishing students who use AI as a partner from those who use it as a shortcut.

A 2025 study of Chinese undergraduates found that students with stronger self-regulation achieved significantly better writing performance in AI-supported contexts (Shi, Liu, & Hu, 2025). Similarly, a large international survey of university students showed that students with higher self-efficacy, intrinsic motivation, and effort regulation were more likely to perceive AI as useful for learning, while students with weaker self-regulated learning tended to adopt AI casually or superficially (Mirriahi et al., 2025) .

In other words, the same tool that tempts one student into bypassing effort can, in the hands of a self-regulated learner, become a collaborator that deepens critical thinking and metacognition.

The Risk of Cognitive Offloading

While there is a world of possibilities that AI brings to the table, we know it comes with a cost. Stadler et al. (2024) found that students using ChatGPT produced arguments that were shallower and less well-reasoned than those of their peers who conducted research via traditional search engines. AI reduces cognitive load, making tasks easier but also diminishing the productive struggle that drives deeper understanding. This is the risk that underlines our core fears about generative AI, cognitive offloading. When the process of thinking can be skipped, learning can feel unnecessary. And when the process feels invisible, students may not see much incentive to plan, explore ideas, revise, or reflect at all. Which is why surfacing that process, and giving it value, is critical.

Is AI Threatening Your Learning Outcomes?
Watch our free, on-demand webinar to get a simple framework for redesigning assignments and turning AI from a shortcut into a cognitive coach.

Moving From Product to Process

The educational technology market has been stuck in a cycle of detection and endless “efficiency”. Detection tools focus on catching AI “plagiarism” after the fact. Productivity tools focus on polishing the final product as quickly as possible. Of course, these approaches are counterproductive. And, in different ways, both bypass the very skills we should be nurturing.

One emerging approach is to shift focus from catching shortcuts to highlighting the process. At Packback, our work on a new feature set, Engagement Insights, is grounded in Zimmerman’s model of self-regulated learning: forethought, performance, and self-reflection. By making these phases visible, we can both encourage and measure the habits that matter most: pausing to plan, persisting through drafting, and reflecting on growth.

This shift matters because the goal is no longer just to detect or deter the one-click shortcut but to equip students with the skills to resist it. Once students realize it’s not just about what they wrote, but how they got there – when students find meaning in the process and know how to engage with it, the hope is that the shortcuts lose their appeal and students gain the ability to thoughtfully use AI to accelerate their learning.

Designing for the Skills That Will Outlast This Moment

A Personal Note

As someone who spent a decade in the classroom and now leads AI product strategy, this challenge is personal. I live the same paradox I described above: I see firsthand the extraordinary potential of AI as an equalizer and accelerator of learning, and yet I feel protective of the magic that happens in the classroom when students wrestle with ideas without automation smoothing the path.

That paradox always brings me to an essential question: How do we preserve the essence of learning while adapting to a world transformed by AI?

My conviction is that AI detection has an expiration date and the path forward is not about searching for AI misuse or merely monitoring students, but about realigning learning around the skills that matter most. By giving students deliberate opportunities for curiosity, self-regulation, reflection, and critical thinking, we can help ensure they don’t just produce work but grow as thinkers. This is about re-centering learning on skills that endure, regardless of how AI evolves.

For educators, this means creating intentional space for that messy thinking, explicitly teaching self-regulation skills, and allowing time for self-reflection to show students that these practices have value. When teachers invest in the process, students learn to invest in it themselves.

And for educational technology companies, it means resisting the pull of fear or hype, and instead being deliberate about where AI can accelerate learning and where timeless skills must remain central to the student experience.

Because at the end of the day, we must believe that students want to learn. The “cognitive time under tension”, the messy, creative struggle of exploring ideas, is what makes learning feel meaningful. If we can nurture that, we not only protect academic integrity in the short term, we also empower students to thrive in a future where AI is everywhere.

About the Author