Google Chirp 3 Voice Model Now Available on Vertex AI

The Rise of Voice in Generative AI

Currently, much of the attention surrounding generative AI centers on text-based interfaces capable of producing text, images, and various other content types. However, a significant shift is occurring, with voice technology emerging as the next major frontier, and its advancement is happening rapidly.

Google recently announced the integration of Chirp 3 – its advanced speech-to-text and high-definition text-to-speech models – into its Vertex AI development platform, with availability beginning next week.

New Voices and Applications

Just last week, Google revealed the rollout of eight new voices for Chirp 3, expanding language support to 31 languages. This platform facilitates the creation of diverse applications, including voice assistants, audiobooks, and automated support agents.

Furthermore, it enables the development of voice-overs for video content. This announcement was made during an event held at Google’s DeepMind facilities in London.

Competition and Innovation in Voice AI

Google’s progress coincides with similar advancements from other companies in the voice AI space. Sesame, the startup behind the remarkably realistic “Maya” and “Miles” AI applications, recently launched its model for developers.

This allows them to build customized applications and services leveraging Sesame’s technology.

Safety Measures and Usage Restrictions

Google is implementing usage restrictions for Chirp 3 to proactively address potential misuse. Thomas Kurian, CEO of Google Cloud, stated that the company is collaborating with its safety team to refine these measures.

Investment and Growth in AI Voice Services

Several major startups, including ElevenLabs, have secured substantial funding – amounting to hundreds of millions of dollars – to accelerate their work in AI voice services.

Chirp 3 and Google’s AI Suite

The integration of Chirp 3 positions it alongside Google’s latest iterations of its flagship large language model, Gemini, which are currently undergoing testing.

It also joins Google’s image-generation model, Imagen, and its sophisticated video generation tool, Veo 2.

Realism and Long-Term Development

Whether Chirp 3 will achieve the same level of “realism” as other AI voice projects, such as the work done by Sesame, remains to be seen.

However, Demis Hassabis, CEO of DeepMind, emphasized that the development of AI is a long-term endeavor, not a quick fix.

The Future of AGI

He doesn’t anticipate artificial general intelligence (AGI) arriving in the immediate future, predicting that it’s still several years away. He believes AI will bring transformative changes over the next decade.

“It’s one of those interesting moments in time,” Hassabis added.

Vertex AI: A Platform for Machine Learning

Google initially launched Vertex AI in 2021 as a platform for developers to construct machine learning services in the cloud.

This launch predated the surge in interest surrounding AI, particularly generative AI, which was sparked by the introduction of OpenAI’s GPT services.

Catching Up and Expanding Capabilities

Since then, Google has been focusing on Vertex AI as it strives to compete with companies like Microsoft and Amazon, which are also developing generative AI tools for developers.

Developers can utilize Vertex AI to classify data, train models, and deploy models for production, in addition to building generative AI applications based on Gemini.

It will be noteworthy to observe whether Google expands Vertex AI to encompass models developed beyond its own internal creations.

The History of Chirp

Google has been developing “Chirp” voice services for many years, initially using the name as a codename for its early efforts to challenge Amazon’s Alexa service.