A not so futuristic scenario: A hospital integrates interconnected medical devices to offer enhanced treatment efficacy, early disease detection, and remote patient monitoring, reducing the workload on healthcare staff and improving patient outcomes. The next level is to assist these devises with cutting-edge AI technologies to ensure system and equipment protection.
Yet, there is an attacker conducting an evasion attack that cause the AI to overlook critical conditions. This could leave healthcare providers and medical data vulnerable, thereby compromising patient safety.
cPAID EU funded project comes to resolve this issue with some simple steps.
Continuous Monitoring:
cPAID strengthens patient security by ongoing surveillance of the AI-assisted medical devices, blocking potential cyber-attacks.
Proactive Security Weakness Resolution:
By incorporating advanced AI driven thread detection & anomaly prevention, cPAID verifies efficient emergency response, facilitating communication and ensuring operational continuity.
Digital Security Awareness:
Using its social engineering module & cyber range, cPAID will elevate cybersecurity knowledge in the healthcare community.
In an era where artificial intelligence is transforming economy, politics and industries, ensuring its security and resilience against cyber threats becomes more critical than ever. The #EUfunded project, cPAID has decided to confront this challenge by developing a holistic defense framework to protect AI systems from adversarial attacks through continuous risk assessment, performing proactive threat mitigation, and robust security practices. cPAID solutions will be validated not only in the Health sector but also in the Energy, Surveillance, Transportation and Cybersecurity Awareness domains.
cPAID, envisions researching, designing, and developing a cloud-based platform-agnostic solution for the comprehensive protection of AI applications and the overall AI operations of organizations against malicious actions. By combining AI-based defense methods, XAI, and Generative AI, among others, this project aims to address both poisoning and evasion adversarial attacks.
We, at Uni Systems, are excited to be at the heart of this initiative as the project coordinator of the cPAID. Emphasizing on innovation, Uni Systems focuses on technological advancements through a dedicated Research and Innovation Unit, that remains engaged in important R&D initiatives, ensuring the company remains a pioneer in emerging technologies. One of the main ways to elaborate innovative ideas and convert them into tangible solutions arises through participation in Horizon projects.
Along with the project coordination comes the establishment of effective communication, control, and reporting infrastructure, while our responsibilities include also the delivery of the Risk Management for AI (RIMA) metamodel, the comprehensive approach to identifying, assessing, and mitigating AI-related risks. The risk identification component will leverage AI-assisted penetration testing specifically designed to evaluate the security vulnerabilities of AI systems by simulating sophisticated adversarial attacks generated through advanced machine learning models.
We are also happy to undertake the AI-based vulnerability analysis to detect weaknesses in AI-driven applications and implement an AI-powered monitoring system, referred to as the AI social watchdog, to identify potential cyber threats. The risk assessment phase will employ AI-based harmonization that will integrate and standardize data from the risk identification processes. Overall, it is planned to be integrated an ongoing risk monitoring with an enhanced notification mechanism able to detect alerts arising from evolving risks, but also to detect new ones, and thus providing complete protection against adversarial AI attacks. The assessment outcomes will propose the respective mitigation actions based on assessment outcomes while the whole process will be continuously assisted by a human-in-the-loop to ensure adaptive and efficient risk management.
Moreover, the involvement of core business and technology partners can significantly strengthen the cPAID’s prospects for successful exploitation. As the project coordinators, we will examine every aspect of the exploitation of the project’s results. The existing distribution channels and networks of the company, can boost the established plans penetrate in the market not only faster but also in a more effective way.
With AI systems facing increasing cybersecurity challenges, the cPAID project is setting new standards for resilience, and Uni Systems is leading the way. As project coordinators and risk management experts we guarantee that AI technologies remain secure against evolving cyberthreats. By incorporating in the project proactive security measures, the project reinforces trust in AI, paving the way for a safer digital future.
Lina GIANNAKANDROPOULOU, PhD | Research Projects Delivery Manager @ UniSystems Luxembourg and cPAID Coordinator
Funded by the European Union. Views and opinions expressed are however those of the author(s) only and do not necessarily reflect those of the European Union or ECCC. Neither the European Union nor the granting authority can be held responsible for them.