Krisp AI Adds Accent Conversion to Meeting Tool
The AI meeting assistant Krisp has expanded its product suite with a new real-time accent conversion feature. The tool is designed to improve clarity and comprehension during live conversations, in addition to its existing noise cancellation and transcription capabilities.
- Krisp's AI-powered accent conversion works in real-time and processes audio directly on the user's device to ensure privacy. The technology is designed to make speakers with various accents more clearly understood, and it offers modes that either provide a different voice profile or preserve the speaker's original voice while reducing their accent. - The company, founded in 2018 by Davit Baghdasaryan and Artavazd Minasyan, initially focused on AI-powered noise cancellation and has a significant research and development team based in Armenia. Krisp's technology is used by major platforms like Discord and RingCentral, processing over 75 billion minutes of audio monthly. - Initially targeting call centers in countries like India and the Philippines to reduce miscommunication and eliminate the need for accent neutralization training, the feature supports a wide range of dialects. This can help companies broaden their talent pool by focusing on skills over accent. - The underlying technology is based on Deep Neural Networks (DNNs) trained on extensive datasets of distinct noises and clean human voices of different genders, ages, and ethnicities. The latest version, AI Accent Conversion v3, is backed by recently awarded patents that focus on enhancing clarity while preserving the speaker's vocal identity. - While Krisp focuses on on-device processing to maintain user privacy, some competitors like Circleback utilize cloud-based AI processing to avoid high CPU usage on the user's machine. Other alternatives in the market include NVIDIA's RTX Voice, which requires specific hardware, and various AI meeting assistants that bundle transcription and summarizing services. - The broader speech-to-speech translation market is projected to grow significantly, with a forecasted size of over $200 million. This growth is driven by advancements in AI and the increasing demand for seamless multilingual communication in business and other sectors. - The rise of AI voice manipulation technologies brings ethical considerations regarding consent, data privacy, and the potential for misuse. Companies in this space are navigating how to ensure users have control over their vocal identity and are transparently informed about how their voice data is used.
