Axians highlights growing threat of deepfakes
IT services company Axians is drawing attention to the increasing risk posed by deepfakes, artificially generated fake video and audio content. Criminals are increasingly exploiting this technology for fraud and disinformation.
IT services and solutions provider Axians is alerting businesses and individuals to the escalating threat of deepfakes, which are fake video and audio materials generated by artificial intelligence (AI). The company warns that criminals are increasingly leveraging this technology for social engineering, bypassing authentication systems, and conducting disinformation campaigns.
The technology allows for the manipulation of faces and voices, presenting individuals in altered contexts or making them appear to say things they never did. Axians illustrates a scenario where an employee makes an urgent money transfer based on a deepfake phone call mimicking their manager, only to learn later that the manager was unaware of the request.
These AI-driven manipulations are employed in various attack vectors. They can include targeted phishing for sensitive information like login credentials or trade secrets, or facilitate "CEO fraud" schemes aimed at illicitly obtaining funds. Furthermore, deepfakes are being used to circumvent biometric security systems, such as voice or video authentication used for remote identification.
Axians emphasizes that while the technology is advancing rapidly, making deepfakes increasingly realistic, some signs of forgery can still be detected with careful examination. The company advises vigilance and suggests looking for anomalies such as unnatural facial expressions, blurry transitions around the face, lack of blinking, or inconsistent video quality. Critically evaluating the source and context of digital media is also crucial.
The underlying technology relies on sophisticated neural networks and deep learning, which analyze vast datasets to mimic human facial features and speech. "Faceswapping" techniques replace faces in videos, while "Face Reenactment" manipulates facial expressions and movements in real-time. Similarly, "Text-to-Speech" and "Voice Conversion" methods are used for audio manipulation. While the technology has creative applications, its malicious use is expected to grow.