- AI voice technology company Eleven Labs is working on deepfakes.
- Although the technology is innovative, it could be misused and has raised concerns from lawmakers.
- The CEO of Eleven Labs said the solution is to digitally watermark synthesized speech.
"Are You a CEO Interested in a Feature Interview?"
All Interviews are 100% FREE of Charge
Companies at the forefront of AI voice technology are grappling with ways to regulate deepfakes without stifling innovation.
“It’s going to be a cat-and-mouse game,” said Matty Staniszewski, co-founder and CEO of ElevenLab. told The Atlantic.
Eleven Lab — Valuation reached $1.1 billion After launching in beta last year, it uses AI to generate compelling audio clips. This includes text-to-speech narration, voice dubbing into 29 languages, and voice cloning. The company claims that a user has generated over 100 voices in the past year.
But lawmakers are concerned that the technology could be misused.
Advances in AI correlate with an increase in high-value phone scams in which identity fraudsters impersonate loved ones, family members, government officials, and more. Biden’s AI chief Bruce Reed even called it “voice cloning” That’s the only thing that keeps him up at night.
And last year, 4chan users abused Celebrities’ tools to generate deepfakes of celebrities spewing racist and transphobic content. According to vice.
But Staniszewski is an idealist.
He believes ElevenLab’s technology will contribute to a world where patients with neurodegenerative diseases such as ALS can communicate vocally even after they have lost the ability to speak. It also has potential as a tool for people to communicate across cultures and languages.
mayor of new york Eric Adams is making robocalls He said Eleven Lab’s technology has allowed him to speak Mandarin, Yiddish and Haitian Creole, allowing him to reach more residents in the city who don’t speak English.
Stanisweski said that to take advantage of this potential while preventing fraud, users need to be able to distinguish between AI-generated voices and human voices. Staniszewski told The Atlantic that the “real solution” is to digitally watermark synthesized voices so humans can tell the real from the fake.
The company is developing this technology, but it will only be effective with the help of other companies. Eleven Lab is accord We are working with several other AI companies, including OpenAI, Anthropic, Google, and Meta, to fight deepfakes in the 2024 election.
Eleven Labs did not immediately respond to Business Insider’s request for comment.