Voice Mode

This is a feature that lets you talk to a chatbot out loud, in real time, instead of typing. Instead of the old relay of "speech recognition → text processing → speech synthesis," the latest models understand sound as sound and respond in sound. That's what makes interrupting mid-sentence, laughter, whispering, and emotionally toned speech possible.

It's drawing attention as a use case for interpretation, pronunciation correction, hands-free use while driving, and as an easy entry point for older adults and children. The movie "Her" is the inevitable comparison people reach for — and it's exactly why debates about "relationships with AI" tend to start with this feature.

1. Open the ChatGPT or Gemini app and tap the voice conversation icon.
2. Try saying, "I'm going to speak English now — correct me if I make a grammar mistake."
3. Interrupt mid-sentence and change the topic — the naturalness of real-time conversation is the whole point of this feature.