On 11 and 12 June, the AIXPERT consortium convened in Athens to initiate a dynamic, 7.5-million euro EU initiative aimed at making artificial intelligence (AI) more transparent, explainable, and robust across a variety of sectors. Partners from eight EU member states, two associated countries (Switzerland and the United Kingdom), and Canada came together in the Hellenic capital and joined online to lay the groundwork for this three-year research and innovation program. The first day featured introductions, an overview of the first year’s work plan, and strategic discussions around deliverables, milestones, ethical considerations, and risk management. Plans for deploying use cases across five key domains – healthcare, recruitment, educational robotics, manufacturing, and the creative industries – were carefully explored, setting the stage for focused action.
Day two dived deep into the technical core of the project, with sessions on AI-agentic systems, evaluation frameworks, explainable foundation model development, and the integration of AI tools. Communication, dissemination, and exploitation strategies were also brought to light. The meeting concluded with a clear roadmap of activities for Year 1, positioning the consortium to make strong strides toward achieving explainable, trustworthy, and human-centered AI.
About AIXPERT
AIXPERT is a Horizon Europe initiative designed to push forward the development of AI that people can trust. At its core, the project seeks to build a flexible backbone for AI systems that can adapt to different contexts while remaining transparent, explainable, and accountable. By combining advances in multi-agent architectures, foundation models, and mechanisms for human feedback, the project aims to create AI that is not only technically robust but also aligned with societal values. To prove its impact, AIXPERT will test its solutions through real-world pilots in fields as varied as healthcare, recruitment, education, manufacturing, and the creative industries, ensuring its approach is practical, sustainable, and beneficial across domains.