2024年3月，数字科学促进与研究协会正式启动"AI治理与伦理"专题研究计划，这是协会首次围绕人工智能社会影响设立的系统性研究项目。该计划由理事刘晓彤主导发起，得到了专家顾问委员会和数智研究院的联合支持。
启动这一计划的背景是人工智能技术在2023至2024年间的爆发式发展。大语言模型和生成式AI的广泛部署带来了前所未有的效率提升，但同时也引发了关于算法偏见、决策透明度、深度伪造、就业替代和数字权利等方面的深切关注。全球多个国家和地区开始加速AI监管立法进程，技术发展与治理框架之间的互动成为数字科学领域最紧迫的议题之一。
"AI治理与伦理"专题研究计划设定了三个核心研究方向。第一，AI决策公平性研究：聚焦算法偏见的检测、度量和消除方法，分析不同应用场景下（如信贷审批、招聘筛选、司法辅助）AI决策的公平性问题及解决方案。
第二，AI治理框架比较研究：系统梳理全球主要经济体的AI治理政策、法规和标准，比较不同治理模式的特点、优势和局限，为务实的AI治理实践提供学术参考。
第三，负责任AI开发实践：研究如何在AI系统的设计、训练、部署和运维全生命周期中嵌入伦理考量，探索可操作的负责任AI开发框架和工具。
计划首批招募了来自AI技术、法学、社会学和公共政策等多个学科背景的研究者参与，体现了ADSPR一贯倡导的跨学科协作理念。研究成果将通过协会年度白皮书、DSRI学术成果库和国际学术会议等渠道发布。
刘晓彤在计划启动会上表示，技术发展不应与社会责任脱节。ADSPR有义务在推动数字科学技术创新的同时，关注技术对人和社会的影响。这一专题研究计划将是协会在数字伦理方向的长期承诺的起点。
In March 2024, the Association for Digital Science Promotion and Research officially launched the special research program on "AI Governance and Ethics". This is the Association's first systematic research project centered on the social impact of artificial intelligence. The program was initiated under the leadership of Council Member Liu Xiaotong and received joint support from the Expert Advisory Committee and the Digital Science Research Institute.
The background for launching this program is the explosive development of artificial intelligence technologies between 2023 and 2024. The widespread deployment of large language models and generative AI has brought unprecedented efficiency gains, but it has also raised deep concerns about algorithmic bias, decision-making transparency, deepfakes, job displacement, and digital rights. Multiple countries and regions around the world have begun accelerating AI regulatory legislation, and the interaction between technological development and governance frameworks has become one of the most urgent issues in digital science.
The special research program on "AI Governance and Ethics" has established three core research directions. First, research on fairness in AI decision-making: focusing on methods for detecting, measuring, and eliminating algorithmic bias, and analyzing fairness issues and solutions for AI decision-making in different application scenarios, such as credit approval, recruitment screening, and judicial assistance.
Second, comparative research on AI governance frameworks: systematically reviewing AI governance policies, regulations, and standards in major global economies; comparing the characteristics, strengths, and limitations of different governance models; and providing academic references for pragmatic AI governance practice.
Third, responsible AI development practice: studying how to embed ethical considerations throughout the full life cycle of AI system design, training, deployment, operation, and maintenance, and exploring actionable frameworks and tools for responsible AI development.
The program's first cohort recruited researchers from multiple disciplinary backgrounds, including AI technology, law, sociology, and public policy, reflecting ADSPR's long-standing commitment to interdisciplinary collaboration. Research outcomes will be released through the Association's annual White Paper, the DSRI academic achievement database, international academic conferences, and other channels.
At the program launch meeting, Liu Xiaotong said that technological development should not become disconnected from social responsibility. ADSPR has an obligation to pay attention to the impact of technology on people and society while promoting innovation in digital science and technology. This special research program will be the starting point of the Association's long-term commitment to digital ethics.