Here's How McAfee's AI-Powered Deepfake Audio/Video Detection Tool 'Project Mockingbird' Will Work

Global digital security tools developer McAfee has announced its deepfake audio detection Project Mockingbird to combat AI voice cloning and deepfake scamming. Revealed at CES 2024, the innovative detection tool will aim to protect its consumers from AI-generated audioscams and deepfake content. McAfee claims that the software has a success rate of over 90% and it can instantly analyze if a piece of content was generated by an AI. Project Mockingbird will be showcased at the CES 2024 in Las Vegas, Nevada.

How Does McAfee’s Mockingbird Work?

Ironically, the deepfake audio detection technology is also an AI-powered solution. The Project Mockingbird technology employs a complex combination of AI-powered contextual, behavioral, and categorical detection models to identify whether the audio, standalone, or video is likely to be AI-generated. The program analyses the context and the behavior of the speaker to determine its legitimacy with a 90% success rate.

The technology will notify users instantaneously if a video from a celebrity doing a promotion, or a presidential candidate spreading misinformation is real or AI-generated for malicious purposes.

“So, much like a weather forecast indicating a 70% chance of rain helps you plan your day, our technology equips you with insights to make educated decisions about whether content is what it appears to be,” said McAfee CTO Steve Grobman in a statement.

Further, according to Grobman, the technology will be officially announced during the CES 2024 exhibition, and the public demos of Project Mockingbird are scheduled to be available onsite at the CES event in Vegas.

The Need for AI-Generated Deepfake Detectors

In 2022, about, one-third (33%) of Americans said they or someone they know had seen or experienced a deepfake scam, as per a Techradar report. Further, according to the U.S. Federal Trade Commission, the new scams which are primarily based on AI technology were targeting and defrauding local old people of their property.

Throughout 2022, older people across the United States lost $1.6 billion to scams, as per the report. Many such scams use AI technology to “clone” the voices of acquaintances and claim that they have been kidnapped to extort money from the victims. They also make use of other tricks that rely on AI generation.

In a report, a committee hearing on AI scams had witnesses who said they received calls that sounded like their loved ones were in danger, injured, or being held hostage. In one of the cases mentioned in the report, an old couple claimed to have received a call from a stranger, and the voice on the phone sounded like their daughter was crying “mom”, asking for help. In another case, Gary Schildhorn, a lawyer from Philadelphia, said at the hearing that the victim almost sent $9,000 to the scammer before his daughter-in-law confirmed it was a racketeering scam.

Further, in the context of elections, public trust in the media, and public figure impersonation, the use of AI deepfakes has become a major concern. The year 2024 is set to be a big one around the globe with elections being held in several countries, including the upcoming US presidential election. Thus, the Project Mockingbird has its work cut out.