如何监管生成式人工智能在科学研究中的应用:风险与策略--基于欧盟和美国人工智能政策的分析

How to Regulate the Application of Generative Artificial Intelligence in Scientific Research:Risks and Strategies--Based on an Analysis of EU and U.S.AI Policies

生成式人工智能因其快速、强大的内容生成能力,在科学研究领域引发广泛关注。如何有效监管其在科学研究中的应用成为讨论焦点。在此方面,欧盟和美国率先采取行动,制定了系列人工智能政策,以全面规范人工智能的研发、部署与使用。基于政策工具和政策内容的二维分析框架对近年来欧盟和美国的人工智能风险管理相关政策进行分析发现:生成式人工智能辅助科学研究可能涉及隐私安全、歧视传播、诚信挑战、监管滞后等风险。欧盟和美国在坚持构建可信赖的人工智能系统的目标定位以及以人为本的价值导向下,通过建立全面的风险管理框架,制定了风险界定与风险应对的具体流程和方法,并通过促进不同利益相关主体的共同参与和合作,明确了各自的权责。同时,为确保监管的顺利实施,欧盟和美国还通过建立数据保护条例、版权保护法律和道德伦理框架等,完善了人工智能科研应用的监管机制。未来,我国应联合各方力量,共同构建细致且适应性强的人工智能应用风险监管框架和相应的政策法律,以确保生成式人工智能在科学研究中的使用合乎道德伦理和法律规范。

Generative artificial intelligence has garnered extensive attention in scientific research due to its rapid and powerful content generation capabilities,sparking discussions on how to effectively regulate its application in scientific research.In this regard,the European Union and the United States have taken the lead in initiating actions,formulating a series of AIGC policies to comprehensively regulate its research,deployment and use.Based on a two-dimensional analytical framework of policy tools and policy content,an analysis of recent EU and U.S.policies related to risk management reveals that the use of AIGC to assist scientific research may pose risks such as privacy and security concerns,dissemination of discrimination,integrity challenges and regulatory lags,etc.The EU and the U.S.,while adhering to the goal of building trustworthy AI and a human-centered value orientation,have established comprehensive risk management frameworks that outline specific processes and methods for risk identification and response.They have clarified the respective rights and responsibilities of different stakeholders by promoting their participation and collaboration.Meanwhile,to ensure the implementation of regulation,the EU and the U.S.have improved the regulatory mechanism by establishing data protection regulations,copyright protection laws and ethical frameworks.In the future,China should unite different parties to construct a detailed and adaptable regulatory framework for AIGC application risks and corresponding policies and laws,ensuring that the use of AIGC in scientific research complies with ethical and legal norms.