Language Science Lunch Talk - Sanshiroh Ogawa

Is (AI-based) Synthetic Speech Ready for L2 Listening Assessment?

Abstract: Research suggests that synthetic speech can be effective in second language (L2) pedagogical contexts, but its application in L2 listening assessment has received less attention. Using synthetic speech in test development could substantially reduce the cost and time required to create listening stimuli. At the same time, prior research highlights unique challenges that synthetic speech may pose, which may or may not be resolved by recent advancements in AI-based systems compared to earlier techniques (e.g., concatenative synthesis). In this talk, I will review the literature underscoring the need to validate synthetic speech in L2 listening assessment and introduce my planned dissertation study, which is currently awaiting data collection.

Lunch served at 12:15 PM.