Skyld

Skyld is an AI model security platform that protects proprietary AI models deployed on edge devices and endpoints. The platform addresses threats related to model extraction, reverse engineering, and unauthorized reuse of AI algorithms. The product offers three main solutions: AI Protection secures on-device AI models against reverse engineering attacks while maintaining low computing overhead. This solution aims to protect intellectual property and competitive advantages embedded in AI models. AI Licensing provides deployment, management, and monetization capabilities for AI models. This solution enables organizations to control how their models are distributed and used while protecting intellectual property rights. AdverScan tests AI model resilience against adversarial examples. This solution allows organizations to evaluate model robustness in controlled environments before production deployment. The platform targets organizations deploying AI models on devices outside their direct control, including smartphones, IoT devices, and vehicles. Skyld addresses security concerns specific to on-device AI implementations where models are vulnerable to extraction and analysis by malicious actors. The company has received recognition through multiple European cybersecurity awards and is backed by partnerships with organizations including Thales, NVIDIA, and Inria.