Table of Contents
What is Audiobox?
Meta’s groundbreaking Audiobox generative AI project, unveiled last month, has taken a leap forward with the launch of a publicly accessible demo. This innovative technology can replicate voices using mere seconds of sample audio, allowing users to create personalized audio samples based on their own voices and text prompts.
Audiobox, the big tech’s latest foundation research model for audio generation, combines voice inputs and natural language text prompts to produce a diverse range of voices and sound effects. The possibilities are vast, from voice descriptions to sound effect generation and audio editing.
The demo showcases a text-to-speech process, featuring only two system voices, “Alice” or “Emily.” Despite the limited options, the technology demonstrates its prowess by translating custom text into alternative audio streams. Users can also enhance their samples by adding custom sounds based on text prompts.
Standout feature
However, and perhaps the most disconcerting, yet standout feature – is the ability to generate your voice. While impressive in accuracy, the potential for misuse raises concerns. Meta acknowledges this by requiring users to agree to a set of terms and conditions before trying this particular function.
Meta charging forward in the AI game
As Meta continues to expand its repertoire of generative AI tools, questions about security and ethical use become paramount. The increasing accessibility of such tools, even with associated terms of service, sparks concerns about potential misuse, especially with elections on the horizon.
Despite reservations about the technology’s readiness for widespread use, Meta is charging ahead with its AI development. The fast-paced nature of advancements in this field is evident, and Meta is determined to stay at the forefront.
Stay updated on all of the latest news by subscribing to the ITP Live newsletter below and by clicking the push notifications.