Jessie Barrot

National University, Philippines

Abstract

This paper argues that generative artificial intelligence (AI) is becoming an active partner in English language teaching (ELT) for lesson planning, communicative practice, feedback, and assessment support. However, classroom use often depends on ad hoc prompts that yield inconsistent pedagogical fit. Because language tasks are shaped by proficiency, skill targets, genre/ register expectations, and sociocultural conditions, ELT needs prompting structures that function as instructional specifications rather than generic technical instructions. In response, the paper proposes CLARIFIES+, a contextualized prompting framework for teaching, learning, and assessment that systematizes ELT-relevant task conditions and boundaries. After reviewing major prompting approaches (e.g., zero-shot, few-shot, chain-of-thought, and contextual prompting) and prominent prompting frameworks, the paper identifies a gap between technical prompt guidance and ELTcentered variables, evaluative alignment, and ethical constraints. CLARIFIES+ addresses this gap through nine core components, namely Context, Limitations, Audience, Role, Intent, Format, Inputs, End product, and Style, plus an adaptive “+” layer for process controls, follow-up clarification, tool use, and safeguards that support accuracy, integrity, and learner safety. The paper illustrates how the framework can support differentiated instruction, learner prompt literacy and self-regulation, and more transparent, consistent AI-supported formative and summative assessment. It concludes by noting limitations, implementation challenges, and future research directions.

Keywords

Contextual prompting, generative AI, prompt engineering, prompting framework, computer assisted language learning