Agentic AI: An Emerging Paradigm of Computing?
Introduction
With the rapid development of large language models (LLMs) and techniques associated with their applications, Agentic AI has been an active topic that attracted both industry and academic communities. The word “Agentic AI” refers to autonomous systems empowered by LLMs to perceive the environment and context, to reflect the changes in the environment, as well as to plan its actions and put the plan into actions to change its designated environment. Unlike traditional generative AI that simply answers prompts, these multi-agent systems interact with external tools and APIs to achieve complex, long-term goals with minimal human supervision. Although autonomous agents and multi-agent systems have been studied intensively in the past three decades, the arrival of LLMs significantly enhanced the intelligence of multi-agent systems. Thus, this opens a new era, which has been widely recognised as revolutionary. A new paradigm of computing is merging.
However, despite the rapid acceptance of agentic AI in both industry and academic communities, many questions in a wide range of perspectives remain. For example,
- From the technology point of view, what are the capabilities and limitations, merits and drawbacks of LLM-empowered agentic AI systems? For example, how to deal with the hallucination problem of LLMs?
- From the methodology perspectives, how should agentic AI application systems should be developed, operated, maintained, and evolved effectively and efficiently? For example, how agentic AI systems should be specified, designed, tested, and their quality and trustworthiness assured?
- From the social perspectives, what are the impacts of LLMs and agentic AI on professional, social, ethic, legal, environmental, and educational aspects? For example, how to ensure the fairness, security, and safety of agentic AI applications?
- And from the philosophy point of view, what kind of world will we live in when the cyberspace is dominated by agentic AI systems that interact and control the robots and physical world? For example, will AGI be achievable and what are the consequences?
This panel provides a forum to active researchers with different backgrounds as the panellists to share their experiences, visions, and concerns on this emerging paradigm of computing, discusses key research questions like the above but not limited to them, and answers questions raised by the audience of the CISOSE congress.
Panellists
Moderator: Prof. Hong Zhu
Oxford Brookes University, UK
Bio: Hong Zhu is a professor of computer science at Oxford Brookes University, UK. His research interests are in software development methodology for cloud-native and AI applications, including software modelling, design, testing and quality assurance, automated software development tools and environments. He has published more than 200 research papers in journals and international conferences. He is a senior member of IEEE, a member of ACM and BSC. He has served on numerous international conferences as general chairs or PC chairs and PC members, including IEEE CISOSE, IEEE COMPSAC, IEEE ICWS, IEEE Cloud, IEEE Edge, etc. He is a co-founder of IEEE AITest and IEEE/ACM AST conferences. He is an associate editor of the Software Quality Journal and the Journal of Multi-Agent and Grid Systems.
Prof. Hiroyuki Sato
National Institute of Informatics, Japan
Bio: Hiroyuki Sato is currently a professor with National Institute of Informatics, Japan. He received the B.Sc., M.Sc., and Ph.D. degrees from the University of Tokyo, Japan, in 1985, 1987, and 1990, respectively. His research interests include computer science and information security. In NII, he is Director of the Center for Trust and Digital Identity Research and Development and represents GakuNin, Japanese academic identity federation. He is a member of IEEE, ACM and IPSJ.
Prof. Jianhua Ma
Hosei University, Japan
Bio: Jianhua Ma is a professor in the Faculty of Computer and Information Sciences, Hosei University, Tokyo, Japan. He was the Director of Hosei University Institute of Integrated Science and Technology in 2024. He served as the Chair of Digital Media Department of Hosei University in 2011-2012. His research interests include pervasive computing, social computing, wearable technology, IoT, smart things, and cyber intelligence. Ma is one of pioneers in research on Hyper World and Cyber World (CW) since 1996, and was a co-initiator of the first international symposium on Cyber World in 2002. He first proposed Ubiquitous Intelligence (UI) towards Smart World (SW), which he envisioned in 2004, and was featured in the European ID People Magazine in 2005. He has conducted several unique CW-related projects including the Cyber Individual (Cyber-I), which was featured by and highlighted on the front page of IEEE Computing Now in 2011. He has published more than 300 papers, co-authored/edited over 15 books and 30 journal special issues, and delivered over 30 keynote speeches at international conferences. He has founded three IEEE Congresses on ‘Smart World’, ‘Cybermatics’ and ‘Cyber Science and Technology’, respectively, as well as IEEE Conferences on Ubiquitous Intelligence and Computing (UIC), Pervasive Intelligence and Computing (PICom), Dependable, Autonomic and Secure Computing (DASC), Cyber Physical and Social Computing (CPSCom), Internet of Things (iThings), Digital Twin, and Metaverse. He is a Chair of IEEE SC Hyper-Intelligence Technical Committee, a Co-chair of IEEE SMC Technical Committee on Cybermatics, and a founding chair of IEEE CIS Technical Committee on Smart World.
Assoc Prof. Tomas Cerny
University of Arizona, USA
Bio: Tomas Cerny is an Associate Professor of Electrical and Computer Engineering at the University of Arizona, Tucson. His research focuses on Cloud Native system evolution, Software Architecture, Technical Debt, Static Analysis, and Software Maintenance. He is the Global Infrastructure Director with the service of 19 years to the International Collegiate Programming Contest Management System. He authored over 200 peer-reviewed publications. Among his awards are the eight best paper awards, the 2026 University of Arizona Mentorship Award, the 2025 Postgrad Awards Most Commented PhD Supervisor of the Year, the 2023 Baylor Scholarship Award, the Outstanding Service Award, the ACM SIGAPP 2018 and 2015 awards, and the 2011 ICPC Joseph S. DeBlasi Outstanding Contribution Award. He actively serves the scientific community as Associate Editor of Cluster Computing Journal and served on the organizing committee for IEEE SOSE, ESOCC, ECSA, SANER, ACM SAC, ACM RACS, and ICITCS.
Assoc. Prof. Andrés Muñoz Ortega
University of Cádiz, Spain
Bio: Andrés Muñoz received his Ph.D. degree in Computer Science from University of Murcia, Spain, in 2011. He is a researcher and Associate Professor at the University of Cádiz, Spain, whose work focuses on Artificial Intelligence, particularly Intelligent Environments and Multi-Agent Systems. His expertise includes context-aware systems, semantic technologies, argumentation-based MAS, ambient assisted living, smart education, smart mobility, and data fusion from physical and social sensors.
Assoc. Prof. Antonio Alberti
University of Leeds, UK
Bio: Antonio Alberti is an Associate Professor in Software Engineering at the University of Leeds, UK. He has more than 20 years of experience in academia and research, with expertise in software engineering, AI, cloud/edge computing, IoT, 5G/6G mobile, DLTs, IoT and Future Internet architectures. He received his PhD from the University of Campinas (Unicamp), Brazil, and was a visiting researcher at ETRI, South Korea. Antonio has published over 130 peer-reviewed papers and has held leadership roles in major research initiatives, including serving as Chief Architect of the NovaGenesis Future Internet project for 18 years and contributing to Brazil’s 6G architecture efforts. His current research focuses on agentic AI, distributed systems, and next-generation network architectures.