McAfee’s new Mockingbird project aims to stop AI voice cloning scams for good

Ahead of CES 2024, McAfee has unveiled a new defensive tool to detect and prevent voice fraudsters and AI-generated cybercriminals from obtaining victims’ sensitive information.

The Mockingbird project, which McAfee claims has a success rate of more than 90%, uses a combination of contextual, behavioral and categorical detection models.Unsurprisingly, Deepfake Audio Detection is itself an AI-powered solution, proving that fighting fire with fire can sometimes be the most appropriate solution.

McAfee unveils the deep voice detector

While such attacks are currently less common, the San Jose, California-based company said early work on the technology comes in anticipation of the rise in audio deepfakes, citing a study that revealed seven in 10 Americans are concerned about deepfakes making… It’s hard to hack. Trust what they see and hear online.

Currently, one-third (33%) of Americans say they or someone they know has witnessed or experienced a deep dive scam. Elections, public trust in the media and public impersonation are major areas of concern for consumers.

Given time – and enough data – scammers may also be able to use deepfakes to impersonate ordinary citizens to gain access to sensitive accounts. Some banks and other institutions use voice recognition-based passwords for phone authentication.

“So, just like a weather forecast showing a 70% chance of rain helps you plan your day, our technology provides you with the insights needed to make informed decisions about whether your content is what it seems,” said Steve Grubman, the company’s CTO.

Public demonstrations of the Mockingbird project are scheduled to be available on site at CES in Las Vegas.

“We aim to give users the clarity and confidence to navigate the nuances of our new AI-driven world, to protect their privacy, identity and well-being online,” Grubman added.

Leave a Reply

Your email address will not be published. Required fields are marked *

Exit mobile version