TrustAudio: Workshop on Secure, Safe, and Privacy-Preserving Speech and Audio AI

Overview

Speech and audio foundation models are increasingly deployed in voice assistants, transcription systems, accessibility tools, and multimodal agents. Both adversarially crafted and naturally occurring audio inputs can induce unintended, unsafe, or harmful behaviors, creating security, safety, and privacy risks.

TrustAudio brings together researchers and practitioners from speech processing, NLP, machine learning, privacy, and security to advance trustworthy and reliable speech and audio AI.