AI & Voice: Consent, Accent, and Who Gets Heard

Voice AI is moving into cars, classrooms, and hiring. It's also a biometric. Join us on consent, accent bias, disclosure, and who gets heard.

Voice is quickly becoming a primary way people interact with AI — in cars, call centers, classrooms, and job interviews. Done well, it removes barriers: people can speak naturally, in their own language, without a keyboard or a form to fill out.

Done poorly, it captures a biometric identifier, misunderstands the very people it was meant to include, and makes consequential judgments about them based partly on
how they sound.

If RMAIIG's August 12th session on AI Voice Technology left you curious about the harder questions underneath the demos, this is where we go deeper.

Join us for a practical, welcoming discussion where we'll explore:

• Voice as biometric: A voiceprint identifies you the way a fingerprint does. What should be captured, retained, or shared, and what does real consent look like?
• Accent and access: Whose speech gets understood on the first try, and who has to repeat themselves? What happens when a system works better for some voices than others?
• Disclosure and trust: Should a voice agent always announce it isn't human? Does the answer change for a patent interview, a classroom, or a career coaching session?
• Intimacy and persuasion: Voice feels personal in a way text doesn't. How does that change trust, attachment, and the potential for manipulation?
• High-stakes conversations: When voice AI mediates access to legal help, teacher feedback, or job opportunities, who benefits and who gets screened out?
• Practical guardrails: What norms would you actually want before deploying a voice system in a school, a workplace, or a public service?

Whether you build voice products, work in education, law, HR, or accessibility, or you're simply curious, bring one moment when a voice system misheard you or heard something it shouldn't have.