Pindrop Security, a voice fraud detection firm, has raised $100 million in debt financing from Hercules Capital to further develop its AI-assisted detection of AI-generated voice deepfakes. This funding adds to the $218.3 million in venture equity funding that Pindrop has secured since its founding in 2011, indicating the company’s stability and growth.
The rise of generative AI, especially after the launch of ChatGPT, has led to an increase in sophisticated voice deepfake attacks. Pindrop reports that such attacks have grown exponentially, with some customers experiencing multiple AI-based deepfake voice attacks daily. These deepfakes can convincingly imitate voices, making it easier for attackers to deceive victims.
Pindrop’s tools detect errors in AI-generated voices by analyzing thousands of audio samples per second, focusing on details like fricatives and timing inconsistencies that are difficult for machines to replicate. CEO Vijay Balasubramaniyan explains how these tiny mistakes help identify fake voices, allowing Pindrop to catch 99% of deepfake voice attacks.
As AI voice deepfakes become more prevalent, Pindrop’s work in voice detection also aids in identifying deepfake videos, as fraudulent videos typically rely on fake voices. By making deepfake detection more challenging for attackers, Pindrop aims to deter such attacks and push bad actors to abandon these tactics.
Reference:
