From October 21 to 23, 2026, Dean Zhu Tianqing and Program Coordinators Ying Zuobin and Liu Chi from the Faculty of Data Science at City University of Macau attended the China National Computer Congress (CNCC 2026) in Chengdu, Sichuan Province. During the conference, Dean Zhu was invited to deliver presentations at two thematic forums, sharing her research insights on cutting-edge topics such as AI agent security and verifiable intelligent collaboration. Organized by the China Computer Federation (CCF), the China National Computer Congress is an annual event for academic and industry exchange in the field of computing. Under the theme “Data-Driven Innovation, Intelligent Computing for the Future,” this year’s congress featured invited talks, plenary forums, and thematic sessions, fostering dialogue between academia and industry on artificial intelligence, data science, and intelligent computing.

CNCC 2026 Conference Backdrop (Left) and Group Photo of Faculty of Data Science Representatives at the Venue (Right)

On October 22, Dean Zhu Tianqing delivered a presentation titled “Emerging Security Challenges for AI Agents” at the thematic forum “AI Ready: Are AI Agents Ready for the Workplace?” The forum focused on the capabilities, collaboration, and trustworthiness of AI agents in real-world work environments.Dean Zhu examined the evolution of AI agents from single-turn question-answering systems to autonomous planning, tool invocation, and multi-agent collaboration, highlighting security challenges such as unintended tool use, unauthorized access, data leakage, and accountability. She further explored technical approaches to establishing trust mechanisms aligned with agent capabilities, levels of autonomy, and task-related risks, including tiered authorization, activity logging, verifiable auditing, and privacy protection. She also emphasized the importance of human intervention and recovery mechanisms, offering research perspectives on the secure deployment of AI agents in practical applications.

Dean Zhu Tianqing Delivering Her Presentation at the “AI Ready” Thematic Forum

On October 23, Dean Zhu Tianqing delivered a presentation titled “From Large Language Models to Autonomous Agents: Building Verifiable Intelligent Collaboration Networks with Blockchain” at the thematic forum “Autonomous but Bound: Adversarial Attacks, Defenses, and Security Governance in AI Agent Systems.” The forum focused on security challenges arising from agent planning, memory, tool invocation, and multi-agent collaboration, exploring approaches to behavioral constraints and security governance.Dean Zhu addressed key issues in the evolution from large language models to autonomous agents, including trusted identity, collaboration incentives, and result verifiability. She introduced approaches to leveraging blockchain technology for identity and behavior verification, as well as incentive and reputation mechanisms to facilitate multi-agent collaboration. She also explored potential applications of these technologies in financial security, privacy computing, and intelligent governance.

Dean Zhu Tianqing Delivering Her Presentation at the “Autonomous yet Bounded” Thematic Forum

The Faculty’s participation in the congress strengthened its academic ties with universities, research institutions, and industry partners in mainland China, while providing valuable opportunities to gain insights into the latest research developments in AI agent security and trustworthy collaboration. The Faculty will continue to actively engage in academic exchanges in data science, artificial intelligence, and cybersecurity, integrating cutting-edge research developments into its efforts to advance academic development and talent cultivation.