Artificial intelligence voice synthesis platform ElevenLabs Inc. today announced it has raised $19 million in fresh funding in a Series A round to launch a new set of tools, including an AI speech detection module.

The ElevenLabs platform allows anyone to produce lifelike speech from text using synthetic voices, cloned voices or completely artificial voices that can be customized according to customer needs. Generated voices can be tailored according to gender, age and accent preferences. Voices can also be set to reflect tone, intonation and emotion, leading to highly realistic content.

The funding round was co-led by Andreessen Horowitz, Nat Friedman and Daniel Gross. Other participants in the round included Credo Ventures and Concept Ventures, and industry angel investors, including Instagram co-founder Mike Krieger, Oculus VR co-founder Brendan Iribe, DeepMind co-founder Mustafa Suleyman and O’Reilly Media founder Tim O’Reilly.

AI voice cloning technology has become a viral sensation for its capability of generating and cloning voices, which makes it an amazing tool for producing speech on demand. According to the company, after spending 2022 researching the voice synthesis technology and rolling out its technology in beta in 2023, it has registered more than 1 million users and generated more than 10 years’ worth of audio content.

However, this has also led to concern about how this technology could be used for nefarious purposes, such as AI deepfakes. One example includes when users of the anonymous English-language image board 4chan began sharing how to use the technology to produce hate speech-laden memes in celebrity voices.

To combat the potential of future deepfake AI voices proliferating, ElevenLabs is unveiling what it calls an AI Speech Classifier. It’s a tool that will allow anyone to determine if what they’re listening to has been generated wholly or in part with ElevenLabs speech synthesis models. Now available to the public, anyone can upload a clip and test voices, as well as select partners as an application programming interface.

ElevenLabs co-founder and Chief Executive Mati Staniszewski said that the release of this tool is part of the company’s commitment to transparency and ethical generative media. “Our mission is to be the ultimate tool for storytelling, dissolving language barriers and putting all audiences in the reach of all content creators in a safe and responsible way,” he said.

Although the tool won’t necessarily stop bad actors from using voice cloning from abusing AI voice technology, it will allow people who are suspicious of voices to test them. Its limitation of only being able to detect ElevenLabs-produced voices could also mean other AI-produced voices will slip through, but it may push other upcoming AI synthesis model makers to create their own.

The company also announced the release of a new product called Projects, a production workflow platform for long-form speech synthesis and audio “infill” that allows users to produce entire dialogue segments for large projects. This includes news articles and even audiobooks – all without leaving the platform.

Using Projects, creators will have full control over directing and editing large project audio including adjusting pacing, inserting pauses and assigning particular speakers to different text fragments. The company said the objective of Projects is essentially to become the “Google Docs” of speech audio generation and editing.

Projects will join ElevenLabs already existing suite of products, including Speech Synthesis, the company’s flagship text-to-speech generation technology, and VoiceLab, which allows users to clone voices or create unique artificial voices for use in their own work.

Image: Pixabay