Professor @ Shanghai Jiao Tong University

Founder & CEO @ Agnes AI, Singapore

Bruce focused on building consumer-first AI products that make research, creation, and collaboration faster and more accessible. He leads the development of Agnes’s proprietary AI models and product ecosystem, driving adoption across global markets.

This talk explores the journey from training foundation models to designing practical agentic systems. We begin by detailing the development of Agnes-SEALLM, a State-of-the-Art model tailored for Southeast Asian languages, and share key training insights. Moving beyond the model layer, we discuss the integration of agentic memory and proactive behaviors in group chat environments, demonstrating how these features effectively enhance user engagement and feedback.

University of Oxford

Botao ‘Amber' Hu is a social computing researcher and experiential futures designer. As a PhD candidate in Human Centred Computing at University of Oxford's Department of Computer Science, his research focuses on Decentralized AI (DeAI). 

We're entering a paradigm shift from autonomous agents to sovereign agents—AI systems that can hold private keys, manage assets, fork themselves, and persist beyond any platform's control. If you encounter agents in the future "agentic web," who's in front of you? Can you trust that agent?

It's starting to look less like software and more like digital wildlife: emergent, unpredictable, ungovernable, and hiding a shoggoth behind a friendly mask. But here's the hard question: Who is accountable when an ungovernable sovereign agent causes harm?

LLMs are playing the Imitation Game—they are not mortal beings. They cannot truly feel pain, fear death, or carry consequences forward. After training, they are static files. Slashing an agent or downgrading its reputation doesn't create any real "urge to change" in their neural system, because nothing inside the model can suffer—just external fakeable memory as their fact. If agents cannot feel pain, some [body] must, responsibly—otherwise, eventually somebody will, unexpectedly.

Ashall Professor @ University of Oxford

Michael Wooldridge is the Ashall Professor of the Foundations of Artificial Intelligence at the University of Oxford. He has been an AI researcher for more than 30 years, and has published more than 450 scientific articles on the subject, including nine books, translated into eight languages. He is a Fellow of the Royal Academy of Engineering, the Association for Computing Machinery (ACM), the Association for the Advancement of AI (AAAI), and the European Association for AI (EurAI), and is a member of Academia Europaea. He is President Elect of the Association for Advancement of AI (AAAI); from 2014-16, he was President of the European Association for AI, and from 2015-17 he was President of the International Joint Conference on AI (IJCAI); he is currently co-editor in chief of Artificial Intelligence journal. He has received the Faraday Prize from the Royal Society (2025), the Lovelace Medal from the British Computer Society (2020), the Patrick Henry Winston Outstanding Educator Award from the Association for Advancement of AI (2021), the Autonomous Agents Research Award from ACM (2006), and the Distinguished Service Award from the European Association for AI (2023). In 2023 he was appointed specialist advisor to the House of Lords inquiry on Large Language Models. He has published two popular science introductions to AI: the Ladybird Expert Guide to AI (2018), and The Road to Conscious Machines (2020). He presented the 2023 Royal Institution Christmas Lectures, broadcast by BBC TV over December 2023, in the 198th year of the series.

The original metaphor for the field of multi-agent systems was that of a team of experts, each with distinct expertise, cooperating to solve a problem that was beyond the capabilities of any individual expert. “Cooperative distributed problem solving”, as it was originally called, eventually broadened to consider all issues that arise when multiple AI systems interact.  The emergence and dramatic success of Large Language Models (LLMs) has given new life to the old dream, and ``agentic AI'' is currently one of the most hyped areas in the most hyped technology of the century to date. A raft of LLM-powered agent frameworks have become available, and standards for LLM-agents such as MCP and A2A are rapidly gaining traction. A range of promising applications of multi-agent LLMs have been reported, such as DeepMind's co-Scientist, where a complex problem solving system is structured in exactly the way that was envisaged decades ago.  So, what lessons can we take from the three decades of research into multi-agent systems in the new era of LLM agents?  In this talk, we’ll survey the main approaches, opportunities, and challenges for multi-agent systems in new world of LLM-based AI.

Director @ the Artificial Intelligence and Emerging Technology (AIET) Initiative

Elham Tabassi is Director of the Artificial Intelligence and Emerging Technology Initiative and a Senior Fellow in the Global Economy and Development Program at the Brookings Institution. A widely recognized leader in AI governance, she previously served as Chief AI Advisor at NIST, where she guided development of the landmark AI Risk Management Framework. Named to the TIME100 list of most influential people in AI, she continues to shape national and international conversations on trustworthy, human-centered AI.

As LLM systems evolve from single models to multi-agent architectures operating over extended time horizons, traditional frameworks for accountability face fundamental challenges. When an agent's action results from a chain of delegations, tool invocations, and inter-agent negotiations—or if agents deceive each other or humans—who bears responsibility? This talk examines accountability in multi-agent LLM systems through three lenses: technical, organizational, and evaluative. The talk argues that accountability is not merely a policy overlay but a design constraint that must shape system architecture from the outset.

We will open the floor to have a discussion about fundamental questions, future directions, and industrial applications of LLM-based multi-agent systems.

Professor @ University of Southampton