Voices shaping the field.

Trustworthy language and multimodal AI: alignment, fairness, robustness and evaluation.

Conversational AI, spoken dialogue, safety, alignment and sociotechnical evaluation.

Safety and security of LLMs and AI agents; organizer of Agents in the Wild.

Trustworthy audio-visual AI, jailbreak attacks and defenses for audio-language models.

Speech and language model security, audio adversarial attacks and over-the-air threats.

Human-centered evaluation of conversational agents: dialogue quality, emotion and empathy.

Adversarial machine learning, trustworthy foundation models, robustness and safety evaluation.

AI agent safety and security, trustworthy machine learning and privacy.