联合国教科文组织发布《教育中的人工智能：可持续发展的机遇和挑战》（附英文版）

人工智能（AI）不断为教学和学习提供新的解决方案，并在全球范围内接受检验。这些解决方案需要先进的基础设施和积极创新的教育生态系统与之配套。这对世界各国，尤其是发展中国家有何影响？人工智能应该成为缩小数字和社会鸿沟的优先事项吗？

教科文组织和布罗孚图卢（ProFuturo）在2019年移动学习周期间发布了《教育中的人工智能：可持续发展的机遇和挑战》工作报告，并在其中探究了这些问题。文件以人工智能技术如何帮助教育系统利用数据推动教育平等、提高教育质量为核心，分析了相关研究案例。

借助中国、巴西、南非等国的具体事例，文件研究了人工智能对学习成果、受教育机会和教师支持的影响。阿联酋、不丹和智利等国的案例则佐证了人工智能如何推动教育管理中的数据分析工作。

此外，文件还探讨了关于人工智能的课程和标准问题，并列举了欧盟、新加坡和韩国的案例，以探讨学生和教师如何为面对充满人工智能的世界做好准备。

在分析机遇的同时，文件也阐述了将人工智能引入教育、帮助学生为人工智能未来做好准备时所遇到的挑战和所收获的政策启示。这些挑战主要在于以下方面：

提升制定全面的人工智能公共政策的能力，助力实现可持续发展：推动该领域发展所需的技术环境十分复杂，这要求多种因素和制度保持和谐一致。务须在国际和国内层面开展公共政策合作，进而建立一个服务于可持续发展的人工智能生态系统。
在教育中应用人工智能时确保包容性和公平性：伴随着人工智能的发展，最不发达国家面临着新的技术、经济和社会分化所带来的种种风险。我们务须直面基本技术基础设施等主要障碍，以创造基本条件、贯彻通过人工智能改善学习的新战略。
帮助教师为人工智能辅助教育做好准备：教师必须掌握新的数字技能，方能以恰当的方式借助人工智能推动教学；人工智能开发人员必须了解教师的工作方式，并创建在现实环境中可持续的解决方案。
发展优质且包容的数据系统：世界正走向教育数据化，数据质量因此应成为主要关注点。提高国家能力进而改善数据的收集和系统化的工作具有关键意义。人工智能的发展应为提高数据在教育系统管理中的重要性提供机会。
加强教育领域人工智能应用的研究：虽然我们能够预见教育领域人工智能的相关研究在未来几年会有所增加，但仍应牢记，教育系统为了促进实践和决策而大力推动教育研究评估时仍存在困难。
处理数据的采集、使用和传播过程中引发的伦理和透明度问题：人工智能在提供教育机会、向学生提供个体化建议、个人数据的集中、责任归属、对工作的影响、数据隐私和数据馈送算法的所有权方面引发了许多伦理问题。实现人工智能监管需要公众围绕伦理、问责、透明度和安全性展开讨论。

2019年移动学习周的关键讨论活动关注这些挑战，为国际教育界、各国政府和其他利益相关者提供了一个独特的机会，共同探讨人工智能给教育的各个领域带来的机遇和威胁。

The challenges and opportunities of Artificial Intelligence in education

rtificial Intelligence (AI) is producing new teaching and learning solutions that are currently being tested globally. These solutions require advanced infrastructures and an ecosystem of thriving innovators. How does that affect countries around the world, and especially developing nations? Should AI be a priority to tackle in order to reduce the digital and social divide?

These are some of the questions explored in a Working Paper entitled‘Artificial Intelligence in Education: Challenges and Opportunities for Sustainable Development’ presented by UNESCO and ProFuturo at Mobile Learning Week 2019. It features cases studies on how AI technology is helping education systems use data to improve educational equity and quality.

Concrete examples from countries such as China, Brazil and South Africa are examined on AI’s contribution to learning outcomes, access to education and teacher support. Case studies from countries including the United Arab Emirates, Bhutan and Chile are presented on how AI is helping with data analytics in education management.

The Paper also explores the curriculum and standards dimension of AI, with examples from the European Union, Singapore and the Republic of Korea on how learners and teachers are preparing for an AI-saturated world.

Beyond the opportunities, the Paper also addresses the challenges and policy implications of introducing AI in education and preparing students for an AI-powered future. The challenges presented revolve around:

Developing a comprehensive view of public policy on AI for sustainable development:The complexity of the technological conditions needed to advance in this field require the alignment of multiple factors and institutions. Public policies have to work in partnership at international and national levels to create an ecosystem of AI that serves sustainable development.
Ensuring inclusion and equity for AI in education: The least developed countries are at risk of suffering new technological, economic and social divides with the development of AI. Some main obstacles such as basic technological infrastructure must be faced to establish the basic conditions for implementing new strategies that take advantage of AI to improve learning.
Preparing teachers for an AI-powered education: Teachers must learn new digital skills to use AI in a pedagogical and meaningful way and AI developers must learn how teachers work and create solutions that are sustainable in real-life environments.
Developing quality and inclusive data systems:If the world is headed towards the datafication of education, the quality of data should be the main chief concern. It´s essential to develop state capabilities to improve data collection and systematization. AI developments should be an opportunity to increase the importance of data in educational system management.
Enhancing research on AI in education:While it can be reasonably expected that research on AI in education will increase in the coming years, it is nevertheless worth recalling the difficulties that the education sector has had in taking stock of educational research in a significant way both for practice and policy-making.
Dealing with ethics and transparency in data collection, use and dissemination: AI opens many ethical concerns regarding access to education system, recommendations to individual students, personal data concentration, liability, impact on work, data privacy and ownership of data feeding algorithms. AI regulation will require public discussion on ethics, accountability, transparency and security.

The key discussions taking place at Mobile Learning Week 2019 address these challenges, offering the international educational community, governments and other stakeholders a unique opportunity to explore together the opportunities and threats of AI in all areas of education.

来源：联合国教科文组织