AuthenticID and Darwinium execs pinpoint AI fraud weaknesses

AI always leaves a trace. Executives from AuthenticID and Darwinium agreed on this point, which offers a silver lining among the storm clouds of AI-enabled fraud they described in a video presentation hosted by Biometric Update.

Advances in AI have democratized fraud, enabled new attack vectors and introduced ways to scale old ones, but also empowered cybersecurity teams with novel defenses and tools.

AuthenticID SVP of Decision Innovation Stephen Thwaits and Darwinium Co-founder and VP of Product Ben Davey discussed emerging fraud attack vectors and tools, and delved into how they work.

Agents are automating fraud attacks, Davey explained, just as they are automating legitimate transactions. Telling which is which is a matter of determining intent.

Gen AI has become too sophisticated to defeat with liveness challenges, and is often indiscernible from the real thing with the naked eye. An example of four videos, one fake and three real, drove home the point, as audience members fared barely better than random chance in detecting which was which (29 percent).

But at the pixel level, AI always leaves a trace, Thwaits assured attendees.  The act of bypassing a device’s camera as part of a biometric injection attack creates telltale signs that make it detectable, if you are looking for the right thing.

Ultimately, Thwaits and Davey gave out some practical advice for how those responsible for their organization’s cybersecurity and fraud protection can understand how to deal with the altered technology landscape through a proactive approach.

Watch the webinar replay on-demand.

The high-stakes duel between AI fraud vs AI detection

Article Topics

AI fraud  |  AuthenticID  |  biometric liveness detection  |  biometrics  |  Darwinium  |  digital ID  |  fraud prevention  |  generative AI  |  injection attacks