Report: How Students and Teachers Worldwide are Adapting to AI

Title

How Students and Teachers Worldwide are Adapting to AI: A Report From the USC Center For Generative AI and Society

Authors

Stephen J. Aguilar
Benjamin D. Nye
William Swartout
Alfredo Macı́as
Yuqing Xing
Rosie Le Xiu

Source

USC Center for Generative Al and Society (via EdArXiv Preprints)
August 2025

DOI: 10.35542/osf.io/wr6n3_v2

Abstract

Introduction

This report presents findings from three complementary studies exploring how generative AI (GenAI) tools are reshaping teaching and learning in higher and K–12 education. Together, these studies highlight how students and educators engage with GenAI tools, what drives or hinders their use, and the opportunities and challenges these technologies present for fostering meaningful learning

Abstract

Generative AI (GenAI) tools like ChatGPT, Claude, and Copilot are transforming how students and educators engage with teaching, learning, and (especially) writing. From personalized feedback to automated content generation, these tools have opened new possibilities about how educational practices are evolving, while also raising important questions about what it means to teach and learn. College students have emerged as early adopters of GenAI, using it to brainstorm ideas and clarify difficult concepts. Unfortunately, they have also used it to outsource their thinking on assignments and act in ways that are academically dishonest. At the same time, educators across the globe are exploring how these tools might streamline their work, enhance instruction, and support student learning. This report presents findings from ongoing work from the USC Center for Generative AI and Society.

The work below examines GenAI’s educational impact from multiple vantage points: student help-seeking behavior, AI-augmented writing support, and international teacher perspectives. Together, these studies show the different ways GenAI is being used in real educational settings.

The first study explores how over 1,000 U.S. college students use GenAI when seeking academic help. It distinguishes between instrumental help-seeking (using AI to understand and learn) and executive help-seeking (using AI to get quick answers). Key findings suggest that students who feel confident in their abilities or have strong internet search skills are less likely to rely on GenAI. Those who avoid peer interaction or trust AI, however, more tend to seek executive help. Notably, when professors encourage thoughtful GenAI use, students are more likely to engage in learning-oriented behaviors.

The second study evaluates ABE (AI for Brainstorming and Editing), a GenAI tool designed to promote reflection and revision during the writing process. Unlike AI systems that generate full essays, ABE guides students through structured, interactive activities such as strengthening arguments, clarifying claims, and considering counterarguments. Students reported that ABE helped them organize their thoughts, broaden their perspectives, and improve their writing—positioning ABE as a “learning companion” rather than a shortcut.

The final study surveyed over 1,500 teachers from five countries (India, the United States, Qatar, Colombia, and the Philippines) to understand how educators are integrating GenAI into their teaching. While many see its value for task automation and personalized learning, concerns about plagiarism, creativity loss, and uneven institutional support remain.

These perspectives highlight the global diversity of GenAI use and the pressing need for professional development and ethical guidance. Taken together, these studies underscore the importance of intentional, reflective integration of GenAI in education—guided by research, design, and policy.

Direct to Full Text Report 

Filed under: News