On 3-4 June 2026, the AIXPERT consortium convened at the Barcelona Supercomputing Center (BSC) in Barcelona, Spain, for its General Assembly and Technical Meeting. Bringing together project partners from across Europe, the two-day event marked an important milestone as the project transitions from foundational research and requirements gathering into the implementation phase of its trustworthy AI platform.
The meeting provided an opportunity to review progress achieved during the first year of the project and to align technical priorities for Year 2. Discussions focused on strengthening collaboration across work packages, validating the technical direction of the AIXPERT architecture, and preparing the deployment of the project’s five pilot use cases.
Opening the meeting, Project Coordinator Haris Papageorgiou (Athena Research Center) highlighted the importance of the coming months for the project. Partners reviewed the achievements of the first year, including the completion of key deliverables related to user requirements, reference architecture, explainability models, ethics compliance, communication and dissemination activities, and exploitation planning. The consortium also discussed the implementation roadmap for the next phase, with a focus on integrating the project’s agentic AI components, multimodal foundation models, governance mechanisms, and evaluation frameworks into a unified platform.
Technical sessions showcased progress across the project’s core research and development activities. Partners presented advances in the development of the AIXPERT explainable multi-agent framework, which aims to enable transparent collaboration between AI agents while making decision-making processes understandable, accountable, and auditable. The consortium also reviewed ongoing work on explainable multimodal foundation models, including mechanisms for model inspection, alignment, explainability, and ethical governance.
A significant part of the meeting was dedicated to the project’s trustworthiness auditing and evaluation framework. Partners discussed methodologies for assessing explainability, transparency, accountability, autonomy, robustness, and fairness of AI systems, ensuring alignment with emerging regulatory requirements such as the EU AI Act. The framework will provide the benchmarks and metrics required to evaluate AI trustworthiness consistently across all pilot domains.
The meeting also provided an opportunity to validate the implementation plans for AIXPERT’s five pilot use cases, spanning healthcare, recruitment, manufacturing, education, and the creative industries. Together, these pilots demonstrate how explainable, trustworthy, and human-centred AI can support complex decision-making in diverse real-world settings. Discussions focused on governance requirements, human oversight mechanisms, transparency and explainability features, ethical and regulatory compliance, and the technical integration of AI agents, knowledge sources, and multimodal models. By reviewing pilot workflows, success criteria, and deployment plans, partners ensured that all use cases remain aligned with the project’s overarching vision of accountable and transparent AI systems that empower rather than replace human decision-makers.
Throughout the meeting, partners worked collaboratively to validate pilot workflows, identify integration dependencies, define evaluation criteria, and address governance, ethics, and data protection considerations. The discussions reinforced AIXPERT’s commitment to developing AI systems that are not only powerful and effective, but also transparent, accountable, and human-centred.
As the project enters its second year, the consortium will focus on implementing the AIXPERT Agentic Patterns Library, developing multimodal explainable AI capabilities, integrating governance mechanisms, and deploying pilot-ready versions of the platform. The outcomes of the Barcelona meeting provide a clear roadmap for achieving these objectives and advancing trustworthy AI solutions across diverse application domains.
The consortium thanks the Barcelona Supercomputing Center for hosting the event and looks forward to the next phase of technical development, integration, and pilot deployment.