June 30, 2026
The Institute of Informatics & Telecommunications (IIT) of NCSR Demokritos is pleased to host an invited talk by Prof. Konstantinos (Kostas) E. Zachariadis, Professor of Financial Economics and Director of Research at the School of Economics and Finance, Queen Mary University of London. Prof. Zachariadis is internationally recognised for his research on market design, game theory, corporate governance, and, more recently, the economics of artificial intelligence and AI-driven markets. The event will take place on Wednesday, 15 July 2026, from 12:00 to 13:00 (EEST) in the Lecture Room of the Institute of Informatics & Telecommunications, NCSR Demokritos, with simultaneous online participation available.
In his lecture, titled “AI Agents Playing Market Games: From Price Discovery to Safe Multi-Agent Trade,” Prof. Zachariadis will explore how AI agents are transforming digital marketplaces and two-sided platforms, from traditional financial markets to emerging markets for compute, energy, data, and AI model components. The talk will examine how market rules can shape the behaviour of strategic AI agents, ensuring efficient price discovery while enabling safe and verifiable multi-agent interactions. Drawing on economic theory, combinatorial auctions, and formal verification methods, the presentation will also discuss future challenges in designing trustworthy AI-native markets.
Registration by 13 July 2026: https://forms.gle/9Gz2U2eR9MwX34XS8
The talk is open to researchers, students, and professionals interested in artificial intelligence, economics, market design, and multi-agent systems.
Bio: Konstantinos (Kostas) E. Zachariadis is an engineer turned economist. He is currently Professor of Financial Economics and Director of Research in the School of Economics and Finance at Queen Mary University of London, a Fellow of the Digital Environment Research Institute (DERI) at QMUL, and a Research Associate at the Financial Markets Group and the Hellenic Observatory Centre at LSE, where he spent the early part of his career as Assistant Professor (Lecturer) in the Department of Finance. His economics and finance research on market design, game theory, corporate governance, and institutional investor behaviour has been published in the Review of Economic Studies, Journal of Finance, Journal of Financial Economics, and Games and Economic Behavior. Earlier work in engineering and operations research has appeared in venues including IEEE Transactions on Communications. Current projects study the economics of the AI wave from different angles: the return on the announced datacenter investment in Greece (Anchoring the Cloud); the design of fair allocation mechanisms for research compute across the planned federated HPC network in the UK (FAIR-Compute); and, more broadly, how market environments — especially for compute resources, models, and data — can be designed and made safe as AI agents become their primary participants. He has extensive teaching experience in derivatives markets, alternative investments, corporate finance, and strategy, spanning undergraduate, postgraduate, executive, and bespoke programmes for private organisations at LSE, QMUL, LBS, and UCL. He holds a PhD in Managerial Economics and Strategy from the Kellogg School of Management at Northwestern University, an MSc in Electrical Engineering and Computer Science from Northwestern, and a five-year Diploma in Electrical and Computer Engineering from the Aristotle University of Thessaloniki. He is a Fulbright Scholar, a recipient of the IEEE Life Members Graduate Study Fellowship, and a Fellow of the Royal Society of Arts (RSA).
Abstract of the talk: Two-sided platforms intermediate much of economic exchange, and will increasingly be populated by AI agents — both inside human markets and in agentnative markets for compute, energy, data, and model components. This talk asks how the rules of such markets can control aggregate behaviour when the participants are strategic and opaque, and whether that control can be verified in the formal sense. Prof. Zachariadis will begin with the established theory for the one-sided, single-period double auction: equilibrium bid-shading vanishes at rate O(1/η), the realised price discovers the fully-revealing competitive price, and the price is asymptotically normal around it. Then he will build a combinatorial double auction for knowledge goods under a No-Assembly constraint, where the platform matches but cannot integrate bundles. Here the clearing linear programme is integral (totally unimodular), and — along any constant-offset equilibrium — bid-shading vanishes and prices concentrate on the competitive price; simulations confirm both rates. The remainder is forward-looking and offered for discussion: a dynamic, Markov-perfect version of the market; institutional norms with formal (Event Calculus / Deontic Logic) semantics that steer equilibria; and the substitution of AI agents for rational traders, which turns control into a measurable, payoff-conditioned quantity and makes verification the binding constraint. Prof. Zachariadis will close with the open theoretical question: verifying norm-governed markets under incomplete information.