Frontier AI company Modulate has raised an additional $25 million from Future Ventures, with participation from Hyperplane and Lakestar, bringing its total funding to $60 million. Modulate creates audio-native AI models that enable machines to understand the nuances of human conversations, whether natural or fake.
It uses a proprietary Ensemble Listening Model (ELM) architecture to generate the individual AI models that power its Velma platform. Velma brings together more than 100 of these specialized models that understand and recognize signals such as emotions, tone of voice, intent, synthetic speech and conversational behavior.
The company currently analyzes more than 10 million hours of audio every month, totaling over 600 million hours. According to Modulate, its transcription and deepfake detection models both ranked #1 on public benchmarks like Hugging Face.
Modulate focuses on using AI to analyze votes rather than generate them. The Velma platform’s results are claimed to be twice as accurate as traditional LLMs and generate seven times fewer false positives. Because Velma works in real time, applications created with it not only understand what is happening, but can also intervene during the process.
Typical applications are used to protect healthcare facilities from deepfake hackers; improve the voice AI agent’s emotion and empathy capabilities; Reduce extremism and harassment on social platforms; Observe and monitor voice agent performance. Detect and stop child care voice conversations; and protect agents from identification in high-risk scenarios with advanced voice masking.
“Language is becoming a primary interface for AI, and that creates a whole new set of problems that can’t be solved using a transcript,” explains Carter Huffman, CEO and co-founder of Modulate. “We are already using audio-native AI to protect companies from deepfake attacks, help voice agents understand emotions and respond more empathetically, detect dangerous behavior in online conversations, and monitor whether voice agents are actually performing as they should.”
The new funding will be used to further develop Modulate’s team and infrastructure and expand the models, APIs, SDKs, integrations and deployment options available to developers and partners building voice applications. while continuing to invest in its own underlying research and technology.
The controversial use of AI-powered fraudulent deepfakes and the simple misuse of speech for abusive purposes are now a fact. Detecting deepfakes and identifying abuse in time to prevent damage is almost impossible without the use of AI. It is difficult to develop applications that can detect misuse of language. The purpose of Modulate is to help with this.
“Developers shouldn’t have to rebuild the audio intelligence layer every time they create a new voice experience,” Huffman added. “Our mission is to build the models and infrastructure that allow them to focus on the application they want to build. The possibilities offered by audio-native AI are expanding incredibly quickly. We have developed the technology and tested it at scale, and this investment allows us to grow the team and move faster to meet this demand.”
Related: EU takes action against AI deepfakes, illegal images and hacking
Related: isVerified escapes stealth with voice recognition apps for deepfakes
Related: Creating deepfakes is becoming easier than ever. Fighting back may require even more AI
Related: The AI arms race: Deepfake generation vs. detection
