Abstract

The rapid proliferation of generative AI tools such as large language models has fundamentally challenged traditional approaches to academic writing and assessment in higher education. This mixed-methods study investigates the perceptions and practices of 543 faculty members and 1,280 students across 15 universities in Europe and Asia regarding the use of generative AI in academic contexts. Survey results reveal that 67% of students have used AI writing tools at least once, while 78% of faculty express concerns about academic integrity. Through thematic analysis of 45 in-depth interviews, the study identifies three dominant paradigms: prohibition, controlled integration, and full adoption. The paper proposes a comprehensive ethical framework—the TRUST model (Transparency, Responsibility, Understanding, Skill-building, and Transformation)—for integrating generative AI into academic writing curricula. The framework has been piloted in four institutions with promising results in maintaining academic standards while leveraging AI capabilities.