Adversarial Clothing and GaP Patches Targeting Clearview AI and Amazon Rekognition
The emergence of Universal Physically Transferable Adversarial Patches (GaP) enables the bypass of black-box facial recognition systems, specifically targeting the computer vision (CV) pipelines used by Clearview AI and Amazon Rekognition. By exploiting vulnerabilities in Convolutional Neural Networks (CNNs) and Transformer-based image classification, GaP patches manipulate physical-to-digital transferability mapping to disrupt feature extraction. This results in significantly higher False Rejection Rates (FRR) and allows users to evade identity matching. The technical vector involves introducing adversarial noise into the physical environment that translates to high-confidence misclassifications within the target model's latent space.
FROST: Malicious Website Exploitation of SSD/NVMe Timing Side-Channels in Google Chrome, Mozilla Firefox, and Brave
The FROST attack is a hardware-level timing side-channel vulnerability that allows malicious websites to conduct high-fidelity user surveillance by analyzing I/O latency signals from NVMe SSD controllers. By leveraging high-resolution browser APIs to measure micro-delays in disk read/write operations, an attacker can fingerprint the specific I/O signatures generated by local applications and operating system processes. This exploitation occurs via a passive "drive-by" mechanism, requiring no user interaction or privileged system access, and fundamentally circumvents existing browser security boundaries, including the Same-Origin Policy (SOP), sandboxing, and privacy-preserving modes such as Incognito or cookie-blocking extensions. Because the signal is derived from shared physical hardware rather than software-defined identifiers, the attack remains invisible to traditional endpoint detection and response (EDR) tools and browser-based privacy mitigations.