Large Model Safety Workshop 2025

Apr 23, 2025, Singapore

In the rapidly evolving landscape of artificial intelligence, the development and deployment of large models have ushered in a new era of innovation and possibilities. However, with great power comes great responsibility, necessitating a deep dive into the safety and ethical considerations surrounding these advanced technologies. To address these critical issues, we are excited to announce a comprehensive workshop focused on Large Model Safety, designed for professionals, researchers, and enthusiasts in the field of AI. This workshop aims to shed light on the myriad challenges and opportunities presented by large models, fostering a robust dialogue on novel attacks, risk mitigation strategies, and ethical guidelines to ensure their beneficial and secure application across various domains.
This engaging workshop will feature a series of expert talks on the complex landscape of large model safety. From understanding the technical intricacies of large model development to exploring the societal impacts of AI safety, attendees will gain insights into the multifaceted aspects of large model safety. Whether you're a seasoned AI practitioner or new to the field, this workshop promises to offer valuable perspectives on fostering innovation while prioritizing safety, ethics, and accountability in the age of transformative AI technologies.

Invited Speakers

maksymandriushchenko

Maksym Andriushchenko

EPFL

mohitbansal

Mohit Bansal

University of North Carolina Chapel Hill

yoshuabengio

Yoshua Bengio

University of Montreal

mysteryspeaker

Bo Li

University of Illinois at Urbana–Champaign

christopherdmanning

Christopher Manning

Stanford University

gagandeepsingh

Gagandeep Singh

University of Illinois Urbana-Champaign

mysteryspeaker

Dawn Song

University of California, Berkeley

xinericwang

Xin (Eric) Wang

UC Santa Cruz

yangzhang

Yang Zhang

CISPA Helmholtz Center for Information Security

Panel Guests

chungchernferng

Chung Chern Ferng

ST Engineering

shameek

Shameek Kundu

AI Verify Foundation

andyta

Andy Ta

Synapxe

danielting

Daniel Ting

SingHealth AI Office

Feedback

QR Code

We’d appreciate it if you could take a few minutes to complete the feedback form to help us improve next year’s workshop.

Program

Registration

Opening Address

Slide
Jun Sun from Singapore Management University

Session 1: Chaired by Jun Sun

Towards Building Safe and Secure AI

Abstract
Dawn Song from University of California, Berkeley

Can We Provide Formal Guarantees for LLM Safety?

Abstract Slide
Gagandeep Singh from University of Illinois, Urbana Champaign

What Is New in LLM Safety? A (Multimodal) Reasoning Perspective

Abstract Slide
Xin (Eric) Wang from UC Santa Cruz

Coffee Break

Session 2: Chaired by David Lo

Large Model Safety: The Narrow Path between Cavalier Building and Paralyzing Fear

Abstract Slide
Christopher Manning from Stanford University

Safety Assessment of Large Generative Models

Abstract Slide
Yang Zhang from CISPA

Trustworthy Planning Agents for Collaborative Reasoning and Multimodal Generation

Abstract
Mohit Bansal from University of North Carolina at Chapel Hill

Lunch

Session 3: Chaired by Jun Sun

Superintelligent Agents Pose Catastrophic Risks: Can Scientist AI Offer a Safer Path?

Abstract Slide
Yoshua Bengio from University of Montreal

Guarding the Future: Advancing Risk Assessment, Safety Alignment, and Guardrail Systems for AI Agents

Abstract
Bo Li from University of Illinois, Urbana Champaign

What Have We Learned From Jailbreaking Frontier LLMs

Abstract Slide
Maksym Andriushchenko from EPFL

Coffee Break

Panel Discussion: Bridging AI Safety and Reliability: When LLMs Meet Real-Life Applications

Abstract
Shameek Kundu from AI Verify Foundation (Moderator)
Chung Chern Ferng from ST Engineering
Andy Ta from Synapxe
Daniel Ting from SingHealth AI Office
Yoshua Bengio from University of Montreal
Gagandeep Singh from University of Illinois Urbana-Champaign
Yang Zhang from CISPA

Closing

Venue

The Grand Ballroom, JW Marriott South Beach, 30 Beach Road, Singapore 189763

The Grand Ballroom at JW Marriott South Beach is housed in a historic Military Drill Hall that dates back to the 1930s, blending heritage with cutting-edge technology. This magnificent venue, capable of accommodating 300+ guests for events, features a 15m by 10m dome-shaped LED wall, creating an immersive visual experience, and is crowned by the stunning Forest of Lights ceiling canopy, a breathtaking display of dynamic lighting effects. As part of a sustainability-focused development designed by Foster + Partners, the venue is Green Mark Platinum-certified, reflecting its commitment to eco-friendly architecture and conservation.

Building Floor 1 Floor 2

Registration

We're full capacity!

Registration for this workshop can be covered by either training or conference budgets.
If you have any questions or seek financial support, kindly approach aidxtech0426@gmail.com or contact@aidxtech.com for help.
Kindly note that the registration is managed by AIDX TECH PTE LTD.
If you need an invitation letter, please fill out the form to apply your invitation letter.

Organizing Committee

General Chair

Speaker 3

Jun Sun

Singapore Management University

Program Co-Chairs

Jingyi Wang
Zhejiang University

Yuqi Chen
ShanghaiTech University

Local Co-Chairs

Peixin Zhang
Singapore Management University

Yifan Jia
AIDX TECH PTE. LTD.

Publicity Chair

James Ong
Artificial Intelligence International Institute

Yedi Zhang
National University of Singapore

Finance Chair

Chris Poskitt
Singapore Management University

Registration Chair

Ling Shi
Nanyang Technological University

Web Chair

Ziyang Feng
AIDX TECH PTE. LTD.

Host & Partners

Host

Partners

Media Support

Supported by