保密信息喂养AI，是保护还是反噬？

Group 1 - The application of trade secret law in the AI field needs to redefine boundaries to alleviate the tension between technological innovation and public interest, with a fundamental premise being the prohibition of using confidential information to train GenAI and other public AI tools [1][11] - The U.S. Patent and Trademark Office and the U.S. Copyright Office have issued guidelines that restrict AI-assisted innovation's patent and copyright protections, emphasizing the need for significant human contribution for patent eligibility and excluding AI-generated materials from copyright protection [1][3] - Trade secret law provides strong protection for AI-related innovations, covering sensitive information that patents and copyrights cannot, such as data used for training AI systems or model weights, making it a vital tool for companies [1][5] Group 2 - The rapid development of AI technology has created a conflict between the insatiable demand for data, algorithms, and computing power and the need for transparency in AI technology, leading to potential societal and ethical impacts [2][7] - Using confidential data to feed GenAI has been deemed illegal in multiple countries, with significant risks of irreversible data leaks, as evidenced by international cases of data breaches involving major companies [3][4] - The systemic flaws in GenAI technology highlight the need for strict regulations, including prohibiting access to core business secrets by public AI and ensuring non-sensitive data undergoes dual processing for identity masking and business context separation [4][6] Group 3 - The expansive application of trade secret law in the AI sector has led to overprotection and potential abuse, resulting in legal, ethical, economic, and social crises that pose systemic challenges [5][6] - The lack of transparency in trade secret law creates a fundamental conflict with the need for accountability in AI technology, particularly in high-risk areas like healthcare and justice, which could lead to severe consequences [6][8] - The concentration of AI technology benefits among a few major players due to trade secret protections raises market barriers and diminishes competition, adversely affecting innovation and resource distribution [7][9] Group 4 - To mitigate the adverse effects of trade secret law in AI, a re-examination of its applicable boundaries is necessary, focusing on balancing corporate rights with public interests through transparency requirements and limited protection scopes [8][11] - The introduction of a compulsory licensing system is essential to prioritize public interest in emergencies, allowing governments to utilize protected technologies while ensuring fair compensation for providers [9][12] - Promoting collaboration between legal frameworks and technology development is crucial for optimizing trade secret law's application in AI, including the integration of transparency requirements into industry standards [9][12]