Research on Secure Sharing of AI-Generated Content Based on Multidimensional Virtual Permutation

• Yifeng Yin,
• Minghui Li,
• Yanhua Zhang,
• Yong Gan,

Vol. 19, No. 6, June 30, 2025
10.3837/tiis.2025.06.010, Download Paper (Free):

• AIGC
• data sharing
• privacy preservation
• Group key agreement
• Multidimensional virtual permutation

Abstract

With the advent of large-scale artificial intelligence (AI) models such as ChatGPT, people have begun to frequently use AI-generated content (AIGC) technology to obtain convenience. However, the training and use of AIGC rely on a large amount of data, and if the data security is not managed properly, there will be user privacy leakage, personal information misuse, and other security issues. The generation of AIGC generally needs to be accomplished by the cooperation of multiple users, which will generate multiple data transfer processes, such as data collection, model training, AIGC distribution, AIGC transaction, etc. All these data sharing processes are accompanied by the flow of user privacy data from multiple entities, which will cause a great threat to user's privacy security if a malicious attacker obtains these data from them. In this paper, we propose a multi-party data security sharing scheme for AIGC based on multidimensional virtual permutation mechanism, which introduces polymorphic timestamps as the initial key sequence, and at the same time enhances the cryptographic strength of the initial key sequence by using self-compilation expansion factors. Without the participation of a trusted third party, multi-party users jointly construct a unique multidimensional virtual permutation function to permute multiple security subsystems, then negotiate the security key by iterating the security subsystems, and finally encrypt the information to achieve secure sharing. The analysis and validation of the scheme show that the scheme is highly secure, can resist most attacks, and has better performance with low computational and communication overheads.

Statistics

Cite this article