Ever wanted to hear Donald Trump speaking Hindi? Try the AI tool that can clone anyone’s voice

By William On May 2, 2023

He has one of the most instantly recognizable voices in Britain, but have you ever wondered what David Attenborough would sound like if he spoke German?

Well, now you can find out, thanks to a new AI tool that can clone someone’s voice and make them say anything in multiple languages.

The tool, from ElevenLabs, requires only a few seconds of audio and even preserves the speaker’s original tone of voice.

Creators hope this will “broaden horizons” in many areas, including publishing, game development, and the media.

You can try it yourself ElfLabs website with your own voice or that of your favorite celebrity!

ElevenLabs software translates voices while preserving the tone of the speaker’s voice

HOW TO TRY THE VOICE CLONER YOURSELF

To try the tool for yourself, visit the Eleven Labs website. Here you can sign up for a free account or pay for more options.

After logging in, a page will appear with options to use a ‘Speech Synthesis’ model.
You can choose from pre-made voices or choose your own voices by clicking ‘Add Voice’.
This will take you to a page called ‘Voice Lab’ where you can design voices from scratch, clone your own voices or ‘clone a voice that you have permission and rights to’.
To clone a sampled recording, click ‘Instant Voice Cloning’ (this is only available for paid accounts).
Subscribed users can then drag and drop/upload any voice recording they have saved on their computer to the site.
Once the voice has been uploaded, you can go back to the ‘Speech Synthesis’ model and click ‘Add Voice’ again.
If you type some words in the text box and click ‘generate’, the voice will read it aloud.

ElevenLabs tweeted, “Introducing Eleven Multilingual v1: our new speech synthesis model! We are excited to launch our new model supporting seven new languages: French, German, Hindi, Italian, Polish, Portuguese and Spanish.

“These advances broaden horizons for creators, game developers and publishers, paving the way for using generative media to create more localized, accessible and imaginative content.”

Twitter was flooded with comments following the announcement, with users testing the software on several familiar faces.

Donald Trump was one of those played with when ElunaAI translated what appeared to be a clone of his voice, saying he had “never caused any controversy.”

Meanwhile, another user changed David Attenborough’s voice to German as the BBC presenter talked about the potential of AI.

“Anyone can do it with this tool,” said the Twitter user. “Also important to note that scammers are definitely going to take advantage of this.”

ElevenLabs claims the model is built to understand the emotion and logic behind words to create natural-sounding and compelling audio.

While it can clone voices from provided samples, the model also has the ability to create completely new voices that have never been spoken before.

Some numbers and symbols can currently be mispronounced, ElevenLabs says, but work is underway to improve them.

The US-based company believes its tool will be useful in a number of different industries, from education to media.

It said: ‘Game developers and publishers can create immersive, localized experiences for international audiences, transcend language barriers and connect with players and listeners to maximize engagement and efficiency, without sacrificing quality or accuracy.

“Educational institutions now have the means to produce audio content for different users in their target languages, improving language comprehension and even pronunciation, and meeting different teaching styles and learning needs.”

Audiobook and content creators were also among those who might find it useful, thanks to the model’s ability to create unique character-like voices.

ElevenLabs added: “Accessibility institutions can now help people with visual impairments or learning disabilities by providing them with tools to easily convert less accessible resources into a medium that suits their needs, both in content and form.”

Creators hope that this new technology will “broaden the horizon” in numerous global industries

The new model can clone voices as well as create entirely new voices that have never been spoken before

Later this year, ElevenLabs hopes to give users full control over the voice created, allowing for tempo adjustments and pauses to be inserted.

“Our panel is here to help you make your vision a reality,” ElevenLabs added.

Despite the benefits, the launch of ElevenLabs comes amid numerous concerns about the wave of new voice cloning technology.

Last month, an anonymous TikTok user, @ghostwriter977, went viral after uploading an AI-generated Drake song featuring The Weeknd.

It raised questions about the gray area of AI-created music and whether it is copyright infringement.

Still, ElevenLabs claims it is “fully committed” to respecting creators’ rights amid its technological advancements.

It said: ‘At Eleven we believe we should strive to make the most of new technologies, but not at all costs.

“As we develop them, we make every effort to implement appropriate safeguards that minimize the risk of harmful misuse. With this in mind, we are fully committed to both respecting intellectual property rights and taking action against misuse.”

MailOnline has reached out to ElevenLabs for comment.