Lingnan University and the University of Bologna co-host International Workshop on Technologies for AI Governance. Experts highlight pathways to stronger AI oversight

The regulatory directions of Agentic AI systems and Open-Source AI have become hot topics globally. In light of this, Lingnan University and the University of Bologna in Italy hosted the 2nd International Workshop on Technologies for AI Governance (TAIG) from 25 to 26 October, as part of the 28th European Conference on Artificial Intelligence (ECAI 2025). The workshop brought together leading experts, academics, and policymakers aiming to improve governance in areas such as AI risk assessment, accountability tracking, and safety assurance, and to foster cross-regional exchange on AI governance between Asia and Europe. The two-day event attracted over 150 academics and students.

Held in a hybrid format both online and on-site at the Faculty of Engineering, University of Bologna, the workshop focused on three core themes: (1) Governance on Open-Source AI (2) Risks and Governance of Agentic AI (3) Technologies Towards Human-Centric AI.

Prof Xin Yao, Vice-President (Research and Innovation) and Tong Tin Sun Chair Professor of Machine Learning at Lingnan University, delivered a keynote speech on Open-Source AI Governance, pointing out that AI is rapidly transforming the world, yet the reasoning behind AI-generated recommendations and decisions remains vague. He emphasised the necessity for the industry to increase the transparency and explainability of AI models through technology, saying “Regulating AI is not merely an ethical or legal issue; it also requires leveraging technology to support monitoring, auditing, and accountability. This will help build safer, more transparent, and governable AI systems. We aim to bring an Asian perspective to the European discussion platform on AI governance, while learning from Europe’s experiences to enhance global blueprints for AI development and applications.”

Dr Takayuki Osogami, Senior Technical Staff Member at IBM Research – Tokyo, noted that certain AI systems can make autonomous decisions, plan ahead, and even devise strategies. Without proper regulation, such agentic AI could pose serious risks, including threats to human safety and global crises. He observed that most existing regulatory efforts focus on computational scale, which is insufficient to gauge actual risk, and proposed assessing the degree of an AI system's autonomy - how many decisions it can make independently - as a more accurate indicator of possible risk than existing metrics that rely on observing environmental states.

Ms Emanuela Girardi, President of the European Association for AI, Data, and Robotics, said that the recent emergence of agentic and physical AI systems acting autonomously in digital and physical environments poses fundamental governance challenges. These technologies operate with unprecedented independence, exceeding traditional regulatory mechanisms designed for human-operated systems. For example, agentic AI can pursue goals along unexpected pathways, potentially causing unintended harm or manipulating users through sophisticated psychological techniques. Physical AI systems, from autonomous vehicles to household robots, present safety concerns in unpredictable human environments where failures can result in physical harm or damage to property. Both the potential for psychological manipulation by AI companions and the deployment of industrial and domestic automated robots demonstrate that governments and institutions are still struggling to keep pace with rapidly evolving new technologies.

Ms Girardi also pointed out that current global AI regulations are fragmented and uncoordinated, as while some countries have approved ethical principles and charters, binding frameworks remain absent. She suggested that frameworks are needed both in Europe and globally to address the unique characteristics of autonomous systems, while ensuring alignment with human values and efficient oversight as technological capabilities accelerate.

Prof Matti Mäntymäki, Professor of Information Systems Science at the University of Turku, Finland, said that industry must translate abstract governance ideals such as regulatory requirements and ethical principles into concrete decision-making and accountability structures. He discussed a five-year collaborative research project with industry, presenting an AI governance framework specifically designed for different organisations whose practical application has provided valuable experience and lessons, offering clear guidance for enterprises to build reliable AI governance capabilities.

The speakers and guests at this workshop were well-known, influential scholars from academia, industry, and governments worldwide, including Prof Xin Yao; Prof Jialin Liu, Associate Professor of the School of Data Science; Prof Christoph Lütge, Full Professor of Business Ethics at the Technical University of Munich; Prof Matti Mäntymäki, Professor of Information Systems Science at the University of Turku; Prof Mirco Musolesi, Professor of Computer Science at the Department of Computer Science at University College London; Prof Fabrizio Riguzzi, Full Professor of Computer Science at the Department of Mathematics and Computer Science of the University of Ferrara; Prof Antonino Rotolo, Professor of legal theory and AI & Law at the University of Bologna; Dr Elliot Mckernon, AI Safety Researcher of Convergence Analysis; Dr Takayuki Osogami, Senior Technical Staff Member of IBM Research – Tokyo; Dr Daniele Proverbio, postdoctoral researcher at the University of Trento; Mr. Tomer Jordi Chaffer, Researcher at McGill University; Ms. Emanuela Girardi, President of the European Association for AI, Data, and Robotics; Ms. Rebekka Görge, Senior Data Scientist of the Fraunhofer Institute for Intelligent Analysis and Information Systems in Germany; and Ms. Anni Lai, Co-Chair of the Generative AI Commons, LF AI & Data.

Please click here for more details about the 2nd International Workshop on Technologies for AI Governance.