Page 312 - Cyber Defense eMagazine September 2025

Public Trust Is Eroding

For enterprise security leaders, it’s important to note just how much of an impact deepfakes have had on
eroding the public’s trust. In a recent global study, 69% of consumers said AI-powered fraud now poses
a greater threat to their personal security than traditional forms of identity theft. As deepfakes and
synthetic media grow more convincing, the line between real and fake continues to blur, and with it,
confidence in online authenticity.

When asked who should be responsible for stopping these threats, 43% of consumers pointed to Big
Tech, while just 18% believed the responsibility lies with themselves. Users have a growing expectation
for enterprises to keep their digital identities safe while engaging with their platform. As manipulated
content becomes harder to detect, service providers must adapt their identity intelligence systems to
meet the growing needs of their customers.

So, what exactly are today’s most pressing identity fraud threats, and how can AI help us detect and stop
them? Let’s break down three of the most concerning threats on the horizon, and the AI-powered identity
intelligence strategies that can stop them in their tracks.

1. Injected Selfies & Deepfakes: Breaking Biometrics

Facial recognition is the foundation of biometric authentication. But what happens when the face isn’t
real?
Today, cybercriminals are increasingly turning to camera injection attacks, which are digitally inserted
images that bypass the physical camera feed. These attacks, sometimes carried out through virtual
cameras, enable fraudsters to simulate a live video stream using pre-rendered or AI-generated faces.
Camera injection attacks are the delivery mechanism behind most deepfake-based fraud, but they’re also
being used as standalone attack vectors.

Deepfake fraud, along with wreaking havoc on social media platforms and tarnishing public figures’
reputations, is now impacting executives in the boardroom. Hackers are leveraging advanced deepfake
tools to highjack high-level executives’ identities and exploit the enterprise. All it takes is a few images, a
few seconds of a digitally recorded voice sample, and an unsuspecting employee to be scammed into
performing costly actions by the fraudster.

One answer to stopping these attacks is the use of liveness detection. This strategy uses AI-based
algorithms that assess the authenticity of the person behind the camera in real time. Advanced systems
can detect micro-movements, light reflections, and other physiological cues that can’t be faked.
Multimodal liveness checks, which combine visual, auditory, and motion-based signals, are quickly
becoming the gold standard.

Cyber Defense eMagazine – September 2025 Edition 312
Copyright © 2025, Cyber Defense Magazine. All rights reserved worldwide.