Don’t miss OpenAI, Chevron, Nvidia, Kaiser Permanente, and Capital One leaders only at VentureBeat Transform 2024. Gain essential insights about GenAI and expand your network at this exclusive three day event. Learn More
A week ago, ElevenLabs, the AI voice startup founded by former Google and Palantir engineers, made headlines with its first major consumer-centric product – a Reader app.
Currently available on iOS, the product is a dedicated voiceover solution that converts any text file or link from the web into AI audio, narrated in different AI voices and accents. Today, the company announced it is expanding this library of voices on the app to include AI voices of late Hollywood celebs Judy Garland, James Dean, Burt Reynolds and Sir Laurence Olivier.
The company has partnered with CMG Worldwide, the firm managing and protecting the intellectual property rights of living and deceased celebrities, to recreate and launch the iconic voices. Additionally, it plans to build on this work with many more celebrated AI voices set to launch in the coming months.
Reader gives AI voice to any digital text
While ElevenLabs has specifically focused on the creative industry with AI models for text-to-speech and speech-to-speech conversion, dubbing and sound effect creation, the Reader app gives a more tailored form to its research in the text-to-speech space. All a user has to do is give the link or file for any digital text – be it an article, PDF, newsletter or 300-page e-book – and the app instantly processes the text and begins the voiceover AI narration, with a green highlighter following along and highlighting each word spoken by the AI.
Countdown to VB Transform 2024
Join enterprise leaders in San Francisco from July 9 to 11 for our flagship AI event. Connect with peers, explore the opportunities and challenges of Generative AI, and learn how to integrate AI applications into your industry. Register Now
The feature is available in English, although users can customize their experience by choosing from 11 voices and accents, from male to female, American to Austrian to British English. Now, the Iconic voices launched today adds to this experience, allowing users to discover and experience content in the voice of the late stars.
Imagine a user being able to listen to L. Frank Baum’s The Wonderful Wizard of Oz in the voice of late Judy Garland who acted in the cinematic adaption of the novel.
For the family members of the late stars, the AI-based voice recreation is an opportunity to make sure that the celebs’ legacies live on, with their existing fans getting a way to reconnect with them, and new-age users getting a way to discover them. Meanwhile, for ElevenLabs, the announcement is expected to drive more engagement on the new app.
“Judy Garland, James Dean, Burt Reynolds and Sir Laurence Olivier are some of the most celebrated actors in history. We deeply respect their legacy and are honored to have their voices as part of our platform,” said Dustin Blank, head of partnerships at ElevenLabs “Adding them to our growing list of narrators marks a major step forward in our mission of making content accessible in any language and voice.”
Are these AI voices safe from abuse?
One of the biggest concerns associated with voice cloning technology – like the one at play here – is that voice recreations of known personalities can portray them as saying things they never actually said in the real world. Biden’s Robocall incident is the biggest example of such an issue. In the same way, what if a CEO’s voice is cloned to make them say things that could potentially ruin their or their company’s reputation?
ElevenLabs says it understands these concerns and is moving to expand partnerships for the iconic voices feature with a particular focus on safety.
Sam Sklar, who handles growth marketing at ElevenLabs, told VentureBeat that the company retains full control over celebrity voices and makes them available only on the Reader app, which has been designed in such a way that users can only convert digital text into AI narration for individual consumption — rather than further sharing or downloading.
“For example, through the Reader App, you could choose an article on VentureBeat and select Judy Garland to narrate it just for you. You cannot access her voice through the ElevenLabs voice library (a separate web product of the company). This means they can’t be used in conjunction with our typical text-to-speech tools on the platform, nor can the content they speak through the Reader App be downloaded or shared,” he explained.
If a user uploads harmful content as text to record its iconic voice narration through a secondary device, the company will not even generate the AI voiceover. It has placed automated and human moderation processes in between to identify and block hate speech and other forms of text that violate its terms of service.
As for the chances of the voice library being misused to clone celeb voices from scratch, Sklar says the platform has been built with several safeguards, including a voice captcha verification that matches the audio samples uploaded for cloning with the voice recording of the user. If the voice doesn’t match after a few attempts, the cloning request is not processed. There’s also a “no go” voices policy in place, which prohibits the cloning of voices deemed high risk.
“Any attempt to clone these voices will be blocked,” Sklar said.
While these steps do reduce the chances of celebs, actors and business executives’ voices being cloned, there still can be cases of violations. For instance, malicious users could craft the content for the Reader app in such a way that it bypasses the moderation measures placed by the company.
In the long run, it will be interesting to see how the iconic voices capability, which has been positioned as an offering for fans and enthusiasts, impacts the industry. The Reader app hosting it will be rolling out both globally and to Android devices this summer. Support for more languages is also on the way.
Source link lol