AI in Education
New research maps how generative AI reshapes student and teacher agency in classrooms
A critical review of 10 studies finds that while generative AI can boost personalized learning, it simultaneously raises concerns about equity and learner autonomy.
Researchers examining the intersection of generative AI and agency in education have identified both promising opportunities and significant risks in how these tools reshape the learning environment.
Key findings from the scoping review#
A new scoping review published on arXiv analyzed 10 academic studies to understand how generative AI affects the ability of students and teachers to make meaningful choices and maintain control over their educational experiences. The team applied a framework rooted in Critical Digital Pedagogy, a lens that emphasizes equity and power dynamics in technology-enabled learning.
The analysis uncovered three dominant themes in the literature:
- Control in Digital Spaces — who decides how AI tools operate and who benefits
- Variable Engagement and Access — inconsistent student participation and uneven availability of these technologies
- Changing Notions of Agency — shifting definitions of what it means for learners to be autonomous actors in AI-enriched classrooms
Promise and peril#
The research suggests generative AI can genuinely enhance student agency by enabling personalized support and adaptive learning paths. However, the same technologies risk narrowing learner autonomy when students become dependent on AI recommendations or when access remains unevenly distributed across socioeconomic groups.
A critical finding is that generative AI may widen existing educational inequalities rather than close them, particularly if deployment lacks intentional design for equitable access.
Implications for policy and practice#
The authors highlight substantial gaps in what researchers currently understand about generative AI's long-term impact on learner agency. They call for educational frameworks that deliberately balance enhanced personalization with safeguards to preserve student choice and control.
The work underscores an emerging consensus among educators: deploying generative AI in schools requires more than technical implementation—it demands careful attention to how these tools affect who has power and voice in the classroom.
