GENERATIVE AI IN EFL ACADEMIC WRITING COURSES: A POLICY AND TASK-DESIGN FRAMEWORK FOR UZBEK UNIVERSITIES
Keywords:
Keywords: generative AI, academic writing, academic integrity, EFL, assessment, task design.Abstract
Abstract. Generative AI tools can produce a fluent English essay in seconds, which undermines the take-home essay as evidence of learning. Responses in Uzbek universities range from silent tolerance to outright bans, and neither helps students learn to write. This thesis proposes a course-level framework with two parts: a three-level AI-use scale assigned to every writing task, with a short disclosure statement, and task-design principles that keep learning visible, including process portfolios, in-class drafting, local topics, critique of AI texts and brief oral defences. It also argues against using detection software as evidence of misconduct.
References
1.Flower L., Hayes J. R. A cognitive process theory of writing // College Composition and Communication. — 1981. — Vol. 32, No. 4. — P. 365–387.
2.Liang W., Yuksekgonul M., Mao Y., Wu E., Zou J. GPT detectors are biased against non-native English writers // Patterns. — 2023. — Vol. 4, No. 7. — Article 100779.
3.Pecorari D. Academic writing and plagiarism: A linguistic analysis. — London: Continuum, 2008.
4.UNESCO. Guidance for generative AI in education and research. — Paris: UNESCO, 2023.