AI in Education Beyond Learning Outcomes: Cognition, Agency, Emotion, and Ethics

AI can harm critical thinking and autonomy: The paper warns that AI-driven "cognitive offloading" (where AI does the thinking for students) and diminished learner agency risk undermining critical thinking, intellectual autonomy, and emotional resilience—capacities essential not just for learning but for democratic participation and informed civic engagement.

Harms extend beyond grades: While AI in education is often evaluated solely on learning outcomes, this research demonstrates that uncritical AI adoption can produce broader societal harms affecting emotional well-being, student agency, and ethical development—dimensions typically overlooked in standard educational metrics.

Multiple problems reinforce each other: The paper identifies how cognitive offloading, reduced agency, emotional disengagement, and surveillance-oriented practices "mutually reinforce one another," creating compounding negative effects rather than isolated issues that can be addressed independently.

Design and governance matter critically: The research emphasizes that AI's impact is not predetermined—"pedagogically aligned, ethically grounded, and human-centered AI systems can scaffold effortful reasoning, support learner agency, and preserve meaningful social interaction," suggesting responsible implementation is possible with intentional choices.

Trust and social connection are at stake: Beyond academic skills, the paper highlights that irresponsible AI integration threatens trust and meaningful human interaction, which are foundational for both effective learning and healthy democratic societies.