Developing an Interpretation Argument for AI-Enhanced Scenario-Based Language Assessments Through a Learning-Oriented Language Assessment Framework

Scenario-based language assessments (SBLAs) measure second and foreign language proficiency through real-world tasks that build critical 21st-century skills. However, their complexity poses challenges to scalability, task authenticity, and validity. This article proposes using the Learning-Oriented Language Assessment (LOLA) framework to guide a partial interpretation argument for integration of artificial intelligence (AI) technologies into a B2 level SBLA. It explores how AI technologies, including generative AI, automated item generation platforms, adaptive feedback systems, and virtual agents can support five LOLA dimensions: proficiency, elicitation, social-interactional, socio-cognitive, and technological. This approach could support the scalability, validity, and pedagogical effectiveness of AI-enhanced SBLA through ongoing validation research.