Resemble AI, a voice cloning platform that utilizes generative AI to create realistic-sounding voices, has recently secured $8 million in a Series A funding round. The investment, led by Javelin Venture Partners and featuring participation from Craft Ventures and Ubiquity Ventures, has raised Resemble’s total funding to $12 million. The company plans to utilize the funds to enhance its enterprise products and expand its team to over 40 employees by the end of the year.
Resemble’s technology has gained significant traction in the media industry, enabling some of the world’s largest media companies to produce previously impossible content. By addressing the challenge of voices not keeping up with frequent version updates in video games, Resemble has positioned itself as a game-changer in the industry. Co-founder and CEO Zohaib Ahmed, who also has experience working at Magic Leap, BlackBerry, and Hipmunk, highlights the impact of Resemble’s technology and its applications beyond gaming.
While Resemble initially focused on gaming use cases, the platform has evolved to offer a range of advanced AI features. These include voice transfer to other languages, personalized messages generated by voice actors, and the creation of real-time conversational agents. However, Resemble is not the only player in the generative voice AI market. Companies such as Papercup, Deepdub, ElevenLabs, Respeecher, Acapela, Voice.ai, and major tech giants like AWS, Azure, and Google Cloud are also prominent players.
The rise of generative voice AI has generated controversy surrounding the rights of voice actors and the potential for malicious use. Voice actors are increasingly being asked to sign away rights to their voices, enabling clients to leverage AI and generate synthetic versions that may eventually replace them. Additionally, malicious actors have used AI to clone people’s voices and deceive victims into thinking they are communicating with someone else. These concerns have raised alarm bells and spurred debates about the ethical implications of this technology.
To address these concerns, Resemble has emphasized its commitment to ethics. The platform enforces strict usage guidelines and requires explicit user consent to clone voices. Users must provide a recording of a consent clip in the voice they intend to clone, ensuring that the voice matches consistently. Resemble also implements measures to prevent misuse during the recording process, flagging potential misuse if users deviate from a predetermined script. Ahmed emphasizes that once voices are created, users own all rights to their voices, and Resemble does not use the data to train other models or sell it to third-party companies.
Resemble has even developed a product called Resemble Detect, which uses AI to authenticate audio data and distinguish real audio from fakes. Resemble Detect analyzes different frequencies to identify artifacts resulting from sound editing or manipulation, providing a confidence prediction regarding the clip’s authenticity. This tool complements Resemble’s audio watermarking technology, PerTh Watermarker, which uses AI to insert imperceptible audio tones carrying identifying information. However, PerTh Watermarker is only compatible with Resemble’s own generated speech.
Resemble’s success is largely attributed to its tools and user base, which has exceeded one million users generating 35 years’ worth of audio in the past 12 months. The company acknowledges the importance of responsible AI use and aims to provide insights and recommendations to ensure the ethical application of generative audio. Resemble’s platform allows content creators to enhance their work by adding authenticity and immersiveness to voice content, offering a new level of engagement for audiences. With the regulation of AI being a top priority for government officials, Resemble’s technology provides a valuable solution for creating high-quality voice content while adhering to ethical standards.
