SURF unveils AI tool to combat growing deepfake threats

SURF Security has announced the launch of the beta version of its AI-powered deepfake detection tool, named the SURF Deepwater deepfake detector, integrated into the SURF Enterprise Zero-Trust Browser.

The tool aims to protect enterprises, media organisations, law enforcement agencies, and military entities worldwide from the growing threat posed by AI deepfakes. It is designed to detect whether a person is a real human or an AI imitation with up to 98% accuracy and swiftly issue alerts about possible deepfake threats.

Recent incidents highlight the financial risks associated with deepfakes, such as a USD $25 million theft involving an AI deepfake impersonation of a CFO. Further, the rise in deepfake technology is also influencing political scenarios, with fake multimedia content targeting high-profile individuals like US President Joe Biden and UK politician Wes Streeting.

Deepfake scams have surfaced globally, expanding rapidly with a 303% increase in the USA and notably higher rates in countries like Portugal, China, and Singapore. Reports of individuals falling victim to AI scams that clone familiar voices further underscore the severity of the issue.

The SURF Deepwater deepfake detector can assess any audio source accessible through the browser, including communications platforms such as Slack, Zoom, and WhatsApp. Users can authenticate audio, either recorded or live, simply with the push of a button.

Ziv Yankowitz, SURF Security's CTO, said, "SURF's Deepwater deepfake detector is the first truly useable, real-time defence against deepfakes. To maximise its effectiveness, we focused on accuracy and speed. The tool's neural network is trained using deepfakes created by the top AI voice cloning platforms, has an integrated background noise reduction feature to clear up audio before processing, and can make a determination in less than 2 seconds. Of course, AI voice cloning software becomes more capable by the day, so like all of cybersecurity, we are committing to winning an ever-evolving arms race."

Swetha Krishnamoorthi, Industry Principal, Cybersecurity at Frost & Sullivan, remarked on the significance of addressing deepfake threats: "The rise of AI-based deepfakes presents significant security challenges for organisations, which can lead to reputation damage, data loss, regulatory non-compliance, and financial losses. SURF Security's AI Voice Detector addresses the challenge right at the user interaction layer - the Browser, which could substantially reduce incident rates without having to invest in yet another deepfake detection tool."

Jarad Carleton, Global Research Director, Cybersecurity at Frost & Sullivan, added, "Deep fakes are in use today by criminal gangs that leverage the technology as part of their confidence scams that attempt to defraud individuals and companies. That's why deepfake detection built into Surf's secure enterprise browser is an exciting and vital feature for enterprise end-users."

SURF Security utilises advanced AI detection technology, employing State Space Models to identify deepfakes across various languages and accents. This is accomplished by examining probabilistic relationships between audio frames for inconsistencies, a method that supports high accuracy and speed, even with brief audio segments.

Future enhancements to the tool are anticipated, including the addition of AI image detection capabilities. SURF intends to collaborate with industry partners to enhance existing open-source databases of deepfake audio and videos. The full release of the product is scheduled for the first quarter of 2025.