Towards Embedding Responsible AI in the School System: Co-Creation with Young People

生成人工智能（GenAI）的最新进展有可能改变教育，从评估实践中的反应性调整到关于我们在（目前狭窄的）机器智能时代应该重视人类教育的基本哲学辩论。虽然现在还为时尚早，但在普遍使用GenAI的时代，对学习的影响是重大的：问责制问题;准确性;和包容性需要解决。年轻人（YP）在人工智能如何能够并且应该用于他们的教育方面有发言权。负责任的人工智能需要与包括YP在内的利益相关者进行有意义的接触，他们有权就影响他们生活的系统进行咨询。该项目将弥合适用于教育人工智能的可解释性、公平性和隐私性原则之间的鸿沟，以及YP在面临其影响尚未完全理解的新兴技术时的价值观、希望和担忧。该项目将与青年教师作为教育界的主要持份者进行生动、具体和有意义的交流，为教育政策和教育实践提出建议。该项目的三个目标是：开发和测试富有想象力、推测性和参与性的方法，以产生对YP对新兴人工智能技术的观点的有意义的见解，在两个不同的教育背景下测试这些方法，并为专注于YP和教育的BRAID示范项目提供强大的方法论基础。为政策制定者提供建议，教育工作者和技术开发人员关于YP认为将GenAI纳入学校学习和评估的重要考虑因素，以及如何培养GenAI素养。为实现这些目标，该项目将：采访学者、主要政府和地方政府以及教育技术（EdTech）公司，了解人工智能和数据目前在苏格兰学校系统中的使用情况，并记录即将到来的变革计划和未来可能的发展。创建教育材料，以培养学习者的GenAI素养，其中包含关于人工智能和数据目前如何在学校中使用的清晰和可访问的视觉摘要，以及关于GenAI在教育环境中的关键新兴想法。与YP小组合作，了解他们对GenAI在教育中可能的、可取的、可接受的未来用途的想法（包括障碍和机会），使用创造性的，投机性的，基于设计和基于故事的方法。传播初步建议，在中学负责任地使用人工智能的决策者，教育工作者和技术开发者。我们将与教育利益相关者网络合作，重新平衡当前参与者的权力和利益。虽然目前政府和地方政府拥有对人工智能和数据在教育中的使用做出决定的权力，但我们的工作将以一种可访问的方式使当前的实践对学习者可见和突出，以便他们能够表达对未来负责任地使用此类技术的知情偏好。这将为学校教育中负责任的人工智能提出行动指导建议，这些建议可以在BRAID计划的后期阶段扩大和颁布。

Recent advances in Generative Artificial Intelligence (GenAI) have the potential to transform education, from reactive tweaks in assessment practices to fundamental philosophical debates about what we should value in the education of humans in an age of (currently narrow) machine intelligence. Though it is still early, the implications for learning in an age of pervasive use of GenAI are significant: issues of accountability; accuracy; and inclusion need addressing. that young people (YP) have a voice in how AI could and should be used in their education. Responsible AI requires meaningful engagement with stakeholders, including YP, who have the right to be consulted about the systems which affect their lives. The project will bridge the divide between principles of explainability, fairness and privacy as they apply to educational AI, and the values, hopes and concerns of YP when faced with emerging technologies whose implications are not yet fully understood. It will produce recommendations for educational policy and visions for educational practice that are grounded in lively, specific and meaningful engagements with YP as key stakeholders in education.The three aims of this project are to:Develop a picture of what responsible GenAI could look like within secondary school education.Develop and test imaginative, speculative and participatory methods for generating meaningful insights into YP's perspectives on emerging AI technologies, testing these methods in two distinct educational contexts and providing a strong methodological foundation for a BRAID demonstrator project focusing on YP and education.Produce recommendations for policymakers, educators and technology developers about what YP consider to be important considerations for including GenAI in school learning and assessment, and how GenAI literacy should be fostered. To achieve these aims, the project will:Interview academics, key government and local government and educational technology (EdTech) companies, to map how AI and data are currently used in the Scottish school system and document upcoming plans for changes and possible future developments.Create educational materials to develop learners' GenAI literacy, containing clear and accessible visual summaries of how AI and data are currently used in schools and key emerging ideas about GenAI in educational contexts.Work with groups of YP to understand their ideas about possible, desirable, acceptable future uses of GenAI in education (including barriers and opportunities), using creative, speculative, design-based and story-based methodologies.Disseminate initial recommendations for the responsible use of AI in secondary schools for policymakers, educators and technology developers.We will engage with a network of educational stakeholders in a way which rebalances the power and interests of current actors. While the power to make decisions about AI and data usage in education currently lies with government and local government, our work will make current practices visible and salient to learners in an accessible way, so that they can express informed preferences about responsible uses of such technology in the future. This will lead to action-guiding recommendations for responsible AI in school education, which can be expanded and enacted in later stages of the BRAID programme.