KDD'25 Tutorial




TrustAgent


A Survey on Trustworthy LLM Agents:




Threats and Countermeasures


With the rapid evolution of Large Language Models (LLMs), LLM-based agents and Multi-agent Systems (MAS) have significantly expanded the capabilities of LLM ecosystems. This evolution stems from empowering LLMs with additional modules such as memory, tools, environment, and even other agents. However, this advancement has also introduced more complex issues of trustworthiness, which previous research focusing solely on LLMs could not cover. In this survey, we propose the TrustAgent framework, a comprehensive study on the trustworthiness of agents, characterized by modular taxonomy, multi-dimensional connotations, and technical implementation. By thoroughly investigating and summarizing newly emerged attacks, defenses, and evaluation methods for agents and MAS, we extend the concept of Trustworthy LLM to the emerging paradigm of Trustworthy Agent. In TrustAgent, we begin by deconstructing and introducing various components of the Agent and MAS. Then, we categorize their trustworthiness into intrinsic (brain, memory, and tool) and extrinsic (user, agent, and environment) aspects. Subsequently, we delineate the multifaceted meanings of trustworthiness and elaborate on the implementation techniques of existing research related to these internal and external modules. Finally, we present our insights and outlook on this domain, aiming to provide guidance for future endeavors.

Detailed Schedule (August xxxth)

TimeSpeakerTitle
11:00 am - 11:10 am Qingsong WenOpening and Introduction
11:10 am - 11:20 am Qingsong WenWhat is LLM Agent?
11:20 am - 12:00 am Qingsong WenTrustAgent Taxonomy
12:00 pm - 13:00 pm - Break
13:00 pm - 14:00 pm Linsey PanIntrinsic Trustworthiness
14:00 pm - 14:40pm Linsey PanExtrinsic Trustworthiness
14:40 pm - 15:00 pm Linsey PanOur Insights
 

Organizers

 

Kun Wang

Postdoctoral Researcher​,
Nanyang Technological University

 

Xinfeng Li

Postdoctoral Researcher​, Nanyang Technological University

 

Yongfeng Zhang

Associate Professor,
Rutgers University

 
 

Linsey Pan

Principal Applied Scientist,
Salesforce

 

Tianlong Chen

Assistant Professor,
The University of North Carolina at Chapel Hill

 

Bo An

President's Chair Professor,
Nanyang Technological University

 
 
 

Qingsong Wen

Head of AI Research & Chief Scientist,
Squirrel Ai

Contributor

 

Miao Yu

Undergraduate,
University of Science and Technology of China

 

Fanci Meng

Graduate, University of Science and Technology of China

 

Xinyun Zhou

Undergraduate,
Zhejiang University

 
 
 

Shilong Wang

Graduate,
University of Science and Technology of China

 

Junyuan Mao

Undergraduate,
University of Science and Technology of China