Text-based AI? That’s so 2024
Matt V
March 18, 2025
AI voice tech is moving fast, and Google is jumping.
Chirp 3, its high-definition voice interface, is landing on Vertex AI next week.
With text and image-based AI already booming, the next big wave is voice, and Google doesn’t want to be left behind.
Last week, Google quietly announced that Chirp 3 is rolling out eight new voices across 31 languages, making it a go-to tool for voice assistance, audiobooks, customer support, and video narration.
The update was revealed at a DeepMind event in London.
But Google isn't the only one making moves:
Sesame, the startup behind the viral AI voices “Maya” and “Miles,” is launching its own developer model for custom apps.
ElevenLabs, one of the biggest AI voice startups, has raised millions to expand its tech.
From chatbots to chatterboxes
To prevent misuse, Google is setting restrictions on Chirp 3, though details are still being figured out.
The model joins Google’s growing AI lineup, sitting alongside the Gemini LLM, Imagen for images, and Veo 2 for video generation.
Of course, not all AI voices are created equal.
While some, like Sesame’s, sound eerily human, DeepMind’s CEO Demis Hassabis say AI is still a work in progress.
He doesn’t believe it’s a “silver bullet” just yet and thinks we’re still years away from artificial general intelligence or AGI.
Google launched Vertex AI in 2021 to help developers build machine learning tools, but with OpenAI’s GPT models dominating the scene, it’s been working hard to catch up to Microsoft and Amazon in Gen AI.
AI voices are getting so good, I might just start ghosting real people.