On September 6, 2023, the inaugural Global AI Defense and Offense Challenge was announced during the 2024 Inclusion Bund Forum, themed “Guarding AI in the Era of Large Models.” This event marks the first specialized technology competition focused on the offense and defense of large AI models in China.
The competition aims to explore practical applications in the AI large model industry by incorporating both offensive and defensive tracks. Participants, including white hat and black hat hackers—programmers who test their own systems for vulnerabilities—will engage in real-world practice on “data poisoning” attacks against generative AI models, as well as competitions for detecting counterfeit content in financial scenarios generated by large models.
The “Offense Track” will emphasize the potential risks associated with the real-world applications of generative AI models. Contestants will employ diverse dynamic attack techniques, such as target hijacking, immersive context insertion, and logical nesting, to provoke the model into generating risky outputs, thereby identifying weakness and enhancing the security resilience of the generative model.
Conversely, the “Defense Track” focuses on detecting fraudulent alterations in financial documents generated by AI, addressing the growing concerns surrounding deepfakes and the risks associated with AI-generated content. Contestants will have access to a training dataset containing over one million altered document samples. They will be tasked with developing and training models, followed by evaluating their effectiveness using the corresponding test set to determine the probability of data forgery.
Jointly organized by the China Society of Image and Graphics, Ant Group, and the Cloud Security Alliance (CSA) Greater China, the competition has received support from prestigious institutions like Shanghai Jiao Tong University, Zhejiang University, and several industry-academic organizations.
The competition features a prize pool of nearly 1 million yuan and has enlisted numerous renowned experts from both academia and industry as judges. Notable figures, including Wang Yaonan, an academician of the Chinese Academy of Engineering and chairman of the China Society of Image and Graphics, and Li Yuhang, president of CSA Greater China, will serve on the expert advisory committee. Nearly 30 distinguished scholars from institutions such as Tsinghua University, Shanghai Jiao Tong University, the Shanghai Artificial Intelligence Laboratory, and the China Society of Image and Graphics will participate in organizing and evaluating the competition.
Registration for the challenge officially opened on September 6, with evaluations set to conclude in early November. Contestants can sign up through the official websites of the China Society of Image and Graphics and Alibaba Cloud’s Tianchi Big Data platform.
