Daon's xSentinel tackles synthetic audio threats through real-time detection

To safeguard against synthetic audio attacks, Daon, the Digital Identity Trust company, has launched xSentinel. This detection system analyses voice channel communication to discern whether an audio input is human or artificially created. The innovative technology aims to protect organisations from the malicious use of deepfakes, voicebots and other AI technologies.

XSentinel, an expansion of Daon's AI.X technology, operates independently of platform and language. The system provides valuable real-time data which organisations can integrate into authorisation workflows, thereby securing their contact centres and voice channels from AI-generated voice attacks.

Technological advancements in deepfake capabilities allow for successfully imitating unique voice signals, including tone and articulation, rendering traditional methods inadequate for detecting voice fraud.

The deepfake technology has caused significant concern in the IT sector. Recent research by Daon revealed that 62% of IT executives in B2B companies expressed apprehension about the security threats posed by AI and deepfakes. The cause for alarm is justified, as numerous Tier One global banks have fallen prey to clones that have bypassed their authentication and security systems.

XSentinel offers a more reliable defence by producing a signal for potential fraud that is more accurate than standard audible cues. Using proprietary algorithms, the tool can detect cues indicative of a digitally created voice just seconds after a caller begins speaking, allowing organisations to incorporate voice protection steps into their processes swiftly.

John Dugan, SVP of South East Asia and ANZ at Daon, stressed the importance of the new technology: "Deepfakes pose a serious threat to the front line of defence against fraudsters, mostly for contact centres. Implementing xSentinel gives businesses the tools to protect themselves and their customers from fraudsters now and in the future."

"This innovative technology is part of Daon's efforts to manage the balance between security and user experience. It detects attacks arising from the misuse of generative artificial intelligence, as part of the AI.X suite of tools for deepfake defence, enabling trust at any time, for any type of transaction, from anywhere."

Tiarnan McCaughan, Senior Market Advisor for Fintech and Financial Services at Enterprise Ireland ANZ, highlighted the local need for such services: "Gartner research showed that the top investment for CIOs in Australia and New Zealand in 2023 was in cyber information and cybersecurity."

"With generative AI, deepfakes, image injection attacks and synthetic voice cloning attacks on the rise in every industry sector, it's never been more important to get ahead of this fraudulent trend."

XSentinel is designed to be a high-performance, user-friendly, affordable solution with real-time signalling for early in-call detection. The tool ensures universal applicability, regardless of language or dialect and can be effortlessly integrated into any contact centre or other voice communication platform. Additionally, the implementation requires no customer enrolment and retains a strong commitment to privacy and compliance, negating the need for biometrics or other PII.