2025 Information Fusion Best Paper Award
A research paper co-authored by Wojciech Samek, Head of the AI Department at Fraunhofer HHI and Professor at the Technical University of Berlin, has received the Information Fusion 2025 Best Paper Award, presented by the leading journal Information Fusion. The award honours the paper “Explainable Artificial Intelligence (XAI) 2.0: A manifesto of open challenges and interdisciplinary research directions”, which sets out a roadmap for future XAI research.
Explainable AI (XAI) has become a cornerstone of trustworthy artificial intelligence, especially as AI systems are increasingly deployed in safety-critical and socially sensitive domains such as healthcare, finance, food systems, and disaster resilience. As the field matures, the paper highlights the limitations of narrowly focused, method-centric approaches to explainability and calls for a broader, more integrated perspective on how explanations are designed, evaluated, and used in practice.
The authors introduce the concept of “XAI 2.0”, emphasising the need for interdisciplinary collaboration, human-centred explanations, robustness, and alignment with societal and regulatory requirements, including the growing role of AI governance. To support this shift, the paper presents a manifesto of 28 open research challenges, grouped into nine categories, ranging from explainability for large language models and generative AI to evaluation, human-centred design, trustworthiness, and societal impact.
The paper brings together 27 leading researchers from various fields across Europe and beyond, offering a comprehensive perspective on the technical, scientific, and societal barriers that still limit the practical uptake of explainable AI. It aims to synchronize research agendas across disciplines and accelerate the translation of XAI research into real-world applications.
The recognition underscores Berlin’s growing role as an international hub for Explainable AI research, supported by institutions such as BIFOLD, one of Germany’s AI competence centres, and by Fraunhofer HHI’s expertise in trustworthy, transparent, and human-centred AI systems. The paper was published in Information Fusion in June 2024 and is available open access.