FAITH at CyberHOT EU: Advancing Trustworthy AI Risk Assessment with TrustGuard

The FAITH project is contributing to discussions on cybersecurity and trustworthy artificial intelligence through two training sessions at CyberHOT EU, focusing on AI risk assessment and cybersecurity risk management. As AI systems become increasingly integrated into critical applications and everyday life, assessing their trustworthiness is essential. Identifying potential risks, protecting AI assets and implementing effective mitigation measures are key steps towards developing and deploying AI systems that are safe, secure and resilient.

TrustGuard: An in-isolation AI Risk Assessment System

13:45–15:15 | Presenters: Eleni Tsalapati and Nicholas Fair

The session TrustGuard: An in-isolation AI Risk Assessment System introduces participants to the Trustworthiness Assessment Framework developed within the FAITH project and offers hands-on training with TrustGuard.

TrustGuard is designed to assess risks associated with compromised trustworthiness characteristics, including safety, security and resilience. Its assessment covers both AI models and data assets throughout the AI lifecycle, from design and development to deployment.

Beyond identifying potential risks, the system provides asset-level controls to support targeted mitigation, helping users understand vulnerabilities and take appropriate action.

Built around a checklist-based approach, TrustGuard operationalises a six-phase, risk-based methodology defined by the FAITH Trustworthiness Assessment Framework. It aligns with ISO 27005 and ISO 42001 and draws on leading European and international guidance and standards, including the EU High-Level Expert Group on AI (HLEG), ENISA, NIST’s AI Risk Management Framework and relevant ISO/IEC initiatives.

The session will give trainees the opportunity to explore how these principles can be applied in practice using Spyderisk – System Security Modeller. No previous experience is required, although familiarity with cybersecurity risk assessment concepts will be beneficial. Participants should bring their own laptops.

Automating Cybersecurity Risk Assessment Through System Modelling

15:30–17:00

The programme continues with Automating Cybersecurity Risk Assessment Through System Modelling, a session exploring the role of system modelling in supporting cybersecurity risk assessment.

Together, the two sessions highlight complementary approaches to identifying, assessing and managing risks in increasingly complex digital environments. While TrustGuard focuses on assessing AI trustworthiness across model and data assets, system modelling offers a structured approach to understanding system components, their relationships and the security risks that may arise from them.

Supporting the Development of Trustworthy AI

Through its work on trustworthiness assessment, FAITH aims to translate principles and standards into practical approaches that support the responsible development and deployment of AI technologies.

The TrustGuard training session provides an opportunity to explore how a structured, risk-based methodology can help identify vulnerabilities, assess potential impacts and guide targeted mitigation measures. By connecting trustworthiness principles with practical assessment tools, the project contributes to wider efforts to make AI systems more transparent, secure, resilient and trustworthy.

More info: https://www.cyberhot.eu/