FAITH Enters Operational Phase: Trustworthy AI Validated Across Seven Large-Scale Pilots

During the second phase of the FAITH project, substantial progress has been achieved as the consortium transitions from framework preparation into full-scale operational deployment across the seven Large-Scale Pilots. Building on the strong foundations established during the first reporting period, FAITH has entered a highly productive stage focused on real-world validation, cross-domain experimentation, and refinement of its human-centric trustworthiness ecosystem. The project remains fully aligned with its Description of Action and amended implementation schedule, with milestones and activities progressing according to plan.

A central milestone of this phase is the operational launch of the first integrated FAITH software suite supporting trustworthiness assessment in practical environments. This includes the deployment of TrustGuard, the core orchestration platform implementing the FAITH AI Trustworthiness Assessment Framework (AI_TAF), together with TrustSense, the maturity and readiness assessment tool for AI teams and stakeholders, and the FAITH AI Model Hub, which incorporates the AI Model Passport and Data Passport functionalities.

These interoperable tools now provide the consortium and pilot owners with a unified environment for analysing risks, documenting AI assets, improving transparency, and strengthening accountability throughout the AI lifecycle. An important achievement of this period is that the seven Large-Scale Pilots have now started actively using the FAITH AI_TAF supporting tools, including the FAITH AI Model Hub, TrustGuard, and TrustSense, as part of their operational workflows and trustworthiness assessment activities. This marks a decisive step from methodology design to practical implementation, allowing pilot partners to evaluate their AI systems under real conditions while systematically documenting assets, identifying vulnerabilities, assessing human readiness, and applying mitigation measures. The practical use of these tools across heterogeneous sectors demonstrates the adaptability and scalability of the FAITH ecosystem.

The seven Large-Scale Pilots are now entering or expanding their execution phases across highly relevant domains: media, transportation, education, underwater robotics, industrial wastewater management, healthcare, and active ageing. In each case, domain-specific AI solutions are being tested under real operational conditions while simultaneously applying the FAITH trustworthiness methodology. This combination of technical deployment and trust evaluation is one of the project’s most distinctive strengths. Pilot partners are generating valuable evidence on fairness, usability, privacy, explainability, robustness, and human oversight, while also identifying sector-specific requirements that will further enhance the FAITH framework.

Strong progress has also been achieved in cross-pilot collaboration and knowledge transfer. Through dedicated workshops, dry-run exercises, bilateral technical meetings, and plenary events, the consortium has established a dynamic mechanism for sharing lessons learned among domains. This ensures that innovations developed in one pilot can inform others accelerating maturity and maximizing impact.

The project’s multidisciplinary nature, combining technical, legal, ethical, societal, and human factors expertise, continues to be a major asset in delivering practical and scalable trustworthy AI solution. Dissemination, communication, and stakeholder engagement activities have continued to expand significantly. FAITH has strengthened its visibility through scientific publications, conference presentations, workshops, webinars, and a continuously growing online presence. Engagement with policymakers, industry, researchers, and end-users is helping position FAITH as an important European reference initiative in trustworthy and human-centric AI. At the same time, exploitation planning and market monitoring activities are advancing to ensure long-term sustainability of the project’s key exploitable results.

Overall, FAITH is progressing nicely and is now demonstrating the concrete value of its methodology in real-world critical domains. The project has moved decisively from conceptual design into operational validation, demonstrating excellent potential to deliver impactful scientific, technological, societal, and policy outcomes during the next stages. By combining innovation with responsibility, FAITH continues to pave the way for trustworthy AI systems that genuinely serve people, organisations, and society.