A new EU-funded project, SHASAI, will address cybersecurity risks in AI systems, from design to real-world operation.
Funded under the Horizon Europe programme, SHASAI aims to strengthen the security, resilience, and trustworthiness of AI-based systems as cybersecurity threats become more common.
It will address cybersecurity risks from the initial design and development stages through to deployment and real-world operation.
Leticia Montalvillo Mendizabal, Cybersecurity Researcher at IKERLAN and SHASAI Project Coordinator, explained: “With SHASAI, we aim to move beyond fragmented security solutions and address AI cybersecurity as a lifecycle challenge.
“By combining secure hardware and software, risk-driven engineering and real-world validation, the project will help organisations deploy AI systems that are not only innovative, but also resilient, trustworthy, and compliant with European regulations.”
Using real-world scenarios to protect AI systems from cybersecurity threats
AI systems are becoming increasingly complex, interconnected, and data-driven, making them targets for cyber-attacks and vulnerable to failures that can have real-world consequences.
The project will demonstrate and validate its methods and tools in three real-world scenarios to help tackle cybersecurity threats:
- AI-enabled cutting machines in the agrifood sector
- Eye-tracking systems used in assistive healthcare technologies
- A tele-operated last-mile delivery vehicle in the mobility sector
These diverse use cases will enable the project’s researchers to test their approach in different fields while ensuring that results can be transferred to other AI applications.
The expected outcome is a robust, adaptive, and trustworthy security architecture that ensures AI systems remain resilient, traceable, and compliant with evolving cybersecurity standards in high-risk environments.
Supporting Europe’s ambitions to promote trustworthy AI
By translating high-level cybersecurity and AI safety principles into concrete technical practices, SHASAI also supports Europe’s broader efforts to promote trustworthy AI.
The project aligns with key EU frameworks and initiatives, including the EU AI Act, the Cyber Resilience Act (CRA), the NIS2 Directive, and the EU Cybersecurity Strategy.
The consortium combines expertise from research organisations, universities, industry, and technology providers. It started on 1 November 2025 and will run until the end of April 2029.
