The project, titled “NLP & AI Agents — LORE: LLM-based Orchestration and AI Agents for Reliable Reasoning,” focuses on how large language models (LLMs) can be organised into collaborating AI agents that reason, verify information, and explain decisions in a trustworthy way. Rather than relying on a single AI model, the team investigates architectures in which an LLM coordinates specialised sub-agents for tasks such as literature retrieval, data analysis, factuality checking, and decision explanation.

From research assistants to clinical and societal applications

The work spans three connected application domains:

• Informed Researcher Assistants, where orchestrated AI agents support scientific reasoning by producing reproducible and verifiable research briefs;
• Clinical Applications, where language-based agents help integrate and summarise complex medical information while reducing hallucinations and increasing accountability;
• Crisis-Resilience Assistants, including tools that monitor and analyse misinformation during crises, strengthening public trust and informed decision-making.

Across these domains, the emphasis lies on transparency, robustness, and human-centred design, with Ethical, Legal and Societal Aspects (ELSA) embedded throughout the research.

Strengthening EWUU collaboration

The AI-Hub team is coordinated by Ayoub Bagheri (UU) and includes researchers from UU (Albert Gatt, Antal van den Bosch) UMC Utrecht (Ruurd Kuiper), TU/e (Joaquin Vanschoren), and WUR (Ricardo Torres), combining expertise in natural language processing, medical AI, multi-agent systems, machine learning, and societal impact assessment. Together, the team contributes directly to EWUU’s themes of Preventive Health and Circular Society by developing AI systems that are not only powerful, but also trustworthy and socially aligned.

Over the coming year, the project will deliver open-source orchestration frameworks, shared evaluation benchmarks, and practical demonstrators. It also serves as a seed initiative for larger national and European research proposals, reinforcing EWUU’s joint position in responsible and explainable AI research.