This interactive webinar is structured into two focused segments that together examine how generative AI is reshaping speaking assessment. Learn how AI is moving us beyond traditional monologic tasks toward authentic dialogic interactions, and discover the evolving nature of speaking assessment constructs.
Sha Liu introduces an eight-layer framework for understanding how AI transforms speaking assessment. Drawing from recent research and British Council initiatives, the session highlights key decision points where speaking constructs may evolve or shift. Get hands-on experience with AI tools and see how design choices impact what is measured in speaking assessments.
Parvaneh Tavakoli examines the complexities of oral fluency in the context of AI-driven assessment. Discover how multiple variables affect fluency, raising important concerns about validity in automated fluency assessment, particularly for dialogic tasks.
Key takeaways:
- Evidence-based insights into current trends in speaking assessment.
- A systematic framework for understanding AI’s impact on speaking constructs.
- Practical AI tool demonstrations and hands-on prompt engineering.
- Protocols for ensuring construct validity in AI-enhanced assessments.
- Ready-to-use templates and prompts for immediate application.
Whether you’re an AI novice or a seasoned expert, this session is designed for diverse technical backgrounds. No prior AI experience required!
Speakers:
- Dr Sha Liu, Test Development Researcher, British Council
- Professor Parvaneh Tavakoli, Department of English Language and Applied Linguistics, University of Reading