人工智能与教育：北京共识（二）

Recommend Actions for Governments and other Stakeholders（Ⅱ）

AI for offering lifelong learning opportunities for all

Reaffirm that the guiding principle for achieving SDG 4 is lifelong learning， which encompasses formal， non-formal and informal learning. Adopt AI platforms and data-based learning analytics as key technologies in building integrated lifelong learning systems to enable personalized learning anytime， anywhere and potentially for anyone， with respect for learners agency. Exploit the potential of AI to enable flexible learning pathways and the accumulation， recognition， certification and transfer of individual learning outcomes.

Be mindful of the need to give appropriate policy attention to the needs of older people， especially older women， and to engage them in developing the values and skills needed for living with AI in order to break the barriers to digital life. Plan and implement well-funded programmes to equip older workers with skills and options that enable them to remain economically active for as long as they choose and to engage in their societies.

Promoting equitable and inclusive use of AI in education

Reaffirm that ensuring inclusion and equity in and through education，and offering lifelong learning opportunities to all， are the cornerstones of achieving SDG 4 – Education 2030. Reaffirm that technological breakthroughs in the field of AI in education are an opportunity to improve access to education for the most vulnerable groups.

Ensure that AI promotes high-quality education and learning opportunities for all， irrespective of gender， disability， social or economic status， ethnic or cultural background， or geographic location. The development and use of AI in education should not deepen the digital divide and must not display bias against any minority or vulnerable groups.

Ensure that AI tools in teaching and learning enable the effective inclusion of students with learning impairments or disabilities and those studying in a language other than their mother tongue.

Gender-equitable AI and AI for gender equality

Underline that the gender gap in digital skills contributes to the low share of women among AI professionals and exacerbates existing gender inequalities.

Affirm our commitment to developing AI applications in education that are free from gender bias and to ensuring that the data used for AI development are gender sensitive. AI applications should drive the promotion of gender equality.

Promote gender equality in the development of AI tools and empower girls and women with AI skills to promote gender equality among AI workforces and employers.

Ensuring ethical， transparent and auditable use of education data and algorithms

Be cognizant that AI applications can impose different kinds of bias that are inherent in the data the technology is trained on and uses as input， as well as in the way that the processes and algorithms are constructed and used. Be cognizant of the dilemmas of balancing between open access to data and data privacy protection. Be mindful of the legal issues and ethical risks related to data ownership， data privacy and data availability for public good. Be mindful of the importance of adopting principles of ethics-， privacy- and security-by-design.

Test and adopt emerging AI technologies and tools for ensuring teachers and learners data privacy protection and data security. Support robust and long-term study of deeper issues of ethics in AI， ensuring AI is used for good and preventing its harmful applications. Develop comprehensive data protection laws and regulatory frameworks to guarantee the ethical， non-discriminatory， equitable， transparent and auditable use and reuse of learners data.

Adjust existing regulatory frameworks or adopt new ones to ensure responsible development and use of AI tools for education and learning. Facilitate research on issues related to AI ethics， data privacy and security， and on concerns about AIs negative impact on human rights and gender equality.

Monitoring， evaluation and research

Be mindful of the lack of systematic studies on the impacts of AI applications in education. Support research， innovation and analysis on the effects of AI on learning practices and learning outcomes， and on the emergence and validation of new forms of learning. Take an interdisciplinary approach to research on AI in education. Encourage cross-national comparative research and collaboration.

Consider the development of monitoring and evaluation mechanisms to measure the impact of AI on education， teaching and learning， in order to provide a valid and robust evidence-based foundation for policy-making.

给政府和其他利益攸关方的行动建议（二）

人工智能服务于提供全民终身学习机会

重申终身学习是实现可持续发展目标4的指导方针，其中包括正规、非正规和非正式学习。采用人工智能平台和基于数据的学习分析等关键技术构建可支持人人皆学、处处能学、时时可学的综合型终身学习体系，同时尊重学习者的能动性。开发人工智能在促进灵活的终身学习途径以及学习结果累积、承认、认证和转移方面的潜力。

意識到需要在政策层面对老年人尤其是老年妇女的需求给予适当关注，并使他们具备人工智能时代生活所需的价值观和技能，以便为数字化生活消除障碍。规划并实施有充足经费支持的项目，使较年长的劳动者具备技能和选择，能够随自己所愿保持在经济上的从业身份并融入社会。

促进教育人工智能应用的公平与包容

重申确保教育领域的包容与公平以及通过教育实现包容与公平，并为所有人提供终身学习机会，是实现可持续发展目标4—2030年教育的基石。重申教育人工智能方面的技术突破应被视为改善最弱势群体受教育机会的一个契机。

确保人工智能促进全民优质教育和学习机会，无论性别、残疾状况、社会和经济条件、民族或文化背景以及地理位置如何。教育人工智能的开发和使用不应加深数字鸿沟，也不能对任何少数群体或弱势群体表现出偏见。

确保教学和学习中的人工智能工具能够有效包容有学习障碍或残疾的学生，以及使用非母语学习的学生。

性别公平的人工智能和应用人工智能促进性别平等

强调数字技能方面的性别差距是人工智能专业人员中女性占比低的原因之一，且進一步加剧了已有的性别不平等现象。

申明我们致力于在教育领域开发不带性别偏见的人工智能应用程序，并确保人工智能开发所使用的数据具有性别敏感性。同时，人工智能应用程序应有利于推动性别平等。

在人工智能工具的开发中促进性别平等，通过提升女童和妇女的人工智能技能增强她们的权能，在人工智能劳动力市场和雇主中推动性别平等。

确保教育数据和算法使用合乎伦理、透明且可审核

认识到人工智能应用程序可能带有不同类型的偏见，这些偏见是训练人工智能技术所使用和输入的数据自身所携带的以及流程和算法的构建和使用方式中所固有的。认识到在数据开放获取和数据隐私保护之间的两难困境。注意到与数据所有权、数据隐私和服务于公共利益的数据可用性相关的法律问题和伦理风险。注意到采纳合乎伦理、注重隐私和通过设计确保安全等原则的重要性。

测试并采用新兴人工智能技术和工具，确保教师和学习者的数据隐私保护和数据安全。支持对人工智能领域深层次伦理问题进行稳妥、长期的研究，确保善用人工智能，防止其有害应用。制定全面的数据保护法规以及监管框架，保证能对学习者的数据进行合乎伦理、非歧视、公平、透明和可审核的使用和重用。

调整现有的监管框架或采用新的监管框架，以确保负责任地开发和使用用于教育和学习的人工智能工具。推动关于人工智能伦理、数据隐私和安全相关问题，以及人工智能对人权和性别平等负面影响等问题的研究。

监测、评估和研究

注意到缺乏有关人工智能应用于教育所产生影响的系统性研究。支持就人工智能对学习实践、学习成果以及对新学习形式的出现和验证产生的影响开展研究、创新和分析。采取跨学科办法研究教育领域的人工智能应用。鼓励跨国比较研究及合作。

考虑开发监测和评估机制，衡量人工智能对教育、教学和学习产生的影响，以便为决策提供可靠和坚实的证据基础。