Interspeech 2026
Abstract: The Interspeech 2026 challenge task on Transfer of Pragmatic Intent in Speech-to-Speech Translation was motivated by the fact that current speech-to-speech translation systems do not prioritize pragmatic fidelity, and thus do not ideally support people conversing across languages. This paper describes the design, datasets, and evaluation methods for the challenge, and the results for the submissions. We find that even the best system lags human performance by 1.2 points on a 5-point scale, and note strengths and weaknesses of our evaluation methods.