AI Startup Tackles Deepfake Fraud with Voice Authentication

(HorizonPost.com) – The voice authentication startup company Pindrop claims to have the technology that could protect millions from losing money to AI-enhanced robocall scams, the Washington Examiner reported.

The increased accessibility of generative artificial intelligence has led to rising concern that scammers could trick people into handing over private information by creating deep fakes that replicate voices with the intent to deceive.

According to Pindrop CEO Vijay Balasubramaniyan, not only are deep fakes used to help spread misinformation, but they could also “break commerce” if businesses are unable to trust that the voice on the other end of a call is a human and not a machine.

Founded in 2015 by Balasabramaniyan, Paul Judge, and Ahamad Mustaque, Pindrop sought to adapt Balasabramaniyan’s 2011 Ph.D. thesis into a marketable product.

In his Ph.D. work at Georgia Institute of Technology, Balasubramaniyan focused on identifying the features in voice calls that could be used to verify the caller’s identity. In his research, he identified several audio characteristics to create a voice fingerprint that could be used to identify whether the voice was authentic.

Since its founding, Pindrop has already engaged in partnerships with several Big Tech giants, including Amazon Web Services and Google Cloud. While the company would not say how many users it currently has, it does claim to work with eight of the top ten US banks and credit unions, three of the top five brokerage firms, and 14 of the top 20 insurance companies.

A company spokesman told the Washington Examiner that Pindrop’s software has analyzed more than 5.3 billion phone calls and helped to prevent $2.3 billion in fraud losses.

While Pindrop’s technology is only in use by commercial companies, Balasubramaniyan said he spoke with several lawmakers while in Washington last week about having the federal government use Pindrop’s services.

Copyright 2023, HorizonPost.com