McAfeeRecently launched a project called "Mockingbird" to deal with the use ofAIThe program uses McAfee's AI deep fake audio detection technology and is named "Mockingbird".
The technology is an innovation McAfee demonstrated at CES 2024 (Consumer Electronics Show). The goal of the technology is to prevent cybercriminals from using manipulated AI-generated audio to commit fraud and manipulate public perception.scamIn a cyberattack, criminals could use a video of a real speaker and manipulate them through deep fake technology to say something that was not actually said, which involves both audio and video.
“Obviously as generative AI develops, we’re seeing cybercriminals, bad actors, quickly turn to using generative AI to build all kinds of scams,” McAfee Chief Technology Officer Steve Grobman said in an interview.
As the election cycle approaches, Grobman noted that they expect to see many forms of generative AI applications in false information as well as legitimate political campaign content generation. As a result, McAfee has increased its investment over the past few years to ensure that it has the right technology for various products and back-end technologies to detect these features and enable users to make more informed judgments on the authenticity of videos.
Grobman stressed that with Mockingbird, they can detect whether audio is authentic, based on analysis of the words being spoken. This is a response to the worrying trend of using generative AI to create realistic deepfakes.
The application of deepfake technology has led to an increase in cybercriminals creating fraudulent content, from voice cloning to imitate loved ones to collude and ask for money to modifying the audio of real videos, making it difficult for consumers to distinguish authenticity in the digital realm.
Grobman pointed out that "Mockingbird" uses a comprehensive application of context, behavior and category detection models provided by AI, with an accuracy rate of more than 90%, which can identify and prevent malicious audio tampering in videos. He likened the technology to a "weather forecast" that helps individuals make wise decisions in digital interactions.
McAfee integrated this deep fake detection technology into its products to protect users from deep fake audio. Given that users are already concerned about being exposed to deep fakes, Grobman pointed out that it is relatively difficult to prevent the spread of deep fake audio on social platforms that are generally considered safe.
Overall, Mockingbird aims to empower users to identify real and fake content to protect against cyberbullying, misinformation, and scams caused by deep fakes. McAfee's goal is to strengthen online privacy, identity security, and overall well-being by providing users with the ability to clearly and confidently discern the authenticity of content in the new AI-based world.
A McAfee survey on deepfake awareness conducted in December 2023 showed that 68% of Americans expressed concerns about deepfakes, of which 33% reported knowledge of or experience with deepfake scams. The main concerns about deepfakes include influencing elections (52%), cyberbullying (44%), undermining credibility in the media (48%), impersonating public figures (49%), creating fake pornographic content (37%), distorting historical facts (43%), and falling into fraud traps where cybercriminals obtain payment or personal information (16%).
Grobman said McAfee builds advanced AI that can recognize tiny features that might be imperceptible even to humans. He noted that while it might not be illegal or unethical to use generative AI to create campaign ads during the political season, consumers will want to know if it was built with generative AI and not based on real audio or video.