This page may contain stale information. Last updated: 2026-09-24
Definition
AI voice refers to neural text-to-speech, voice cloning, and speech-to-speech models used as creator tools and as a primary interface for conversational agents and enterprise apps.
Key Points
-
2026-09-23: google gemini-38-flash-tts — generative voice design from natural language, 100+ languages, consent-gated 30s voice replication, SynthID watermarking (2026-09-23-gemini-3-8-text-to-speech)
-
2026-09-23: openai extends gpt-live to ChatGPT mobile Work tab for Plus/Pro — voice-triggered agentic workflows (2026-09-23-chatgpt-mobile-voice-agentic-features)
-
2026-09-15: gemini-38-live Extended Thinking ranks #1 on Artificial Analysis Speech-to-Speech index (82.6); 97-language mid-conversation switching (2026-09-15-gemini-3-8-live-extended-thinking)
-
2026-07-31: smallest-ai $13M Series A for ultra-low-latency full-duplex-voice enterprise agents (2026-07-31-smallest-ai-13m-series-a)
-
Open-source TTS projects often seed commercial APIs (Fish Speech → Fish Audio)
-
Consent, DMCA, and voice-likeness rights are recurring product/legal risks
-
Competitive landscape includes elevenlabs and multilingual low-latency APIs