AI + Dance: Co-Designing Culturally Sustaining Curricular Resources for AI and Ethics Education Through Artistic Computing
AI + Dance: Co-Designing Culturally Sustaining Curricular Resources for AI and Ethics Education Through Artistic Computing
复制标题
人工智能舞蹈：通过艺术计算共同设计人工智能和道德教育的文化可持续课程资源
DOI：
--
复制
发表时间：
2022
期刊：
影响因子：
--
通讯作者：
K. McDermott
中科院分区：
文献类型：
--
作者：
F. Castro;Kayla Desportes;W. Payne;Yoav Bergner;K. McDermott
人工智能(AI)和机器学习(ML)系统在许多领域无处不在，从医学(例如，肿瘤检测)到自然语言处理(例如，数字家庭助理、自动翻译工具)，以及个性化(例如，社交媒体推荐)。随着这些系统的兴起，其中一些系统的设计和实施方式带来的道德挑战也有所增加。最有问题的是它们的传播和放大，如种族主义和性别歧视[4，6]等问题。此外，这些挑战存在于一个计算学科中，该学科已经被排外的、边缘化的文化和实践所拖累，导致妇女和黑人、土著和有色人种的参与率较低[3]。我们的项目，AI+Dance，通过发展我们对如何使学习者能够在包容和文化持续的体验中认识和纠正AI和ML系统的问题的理解，关注AI/ML教育中的这些不平等方面。我们与STEM from Dance1(SFD)合作探索这一点，SFD是一个非营利性组织，支持有色人种女孩通过舞蹈、CS和STEM进行创意制作。在我们之前与SFD的合作中，我们开发了DanceOn，这是一个开放访问的创造性编码环境，使学习者能够创建代码来真实地参与舞蹈和身体动作[5]。有了DanceOn，学习者可以编写可以绑定和响应身体位置的动画，并在空间中静态和动态定位。该系统提供了两种探索AI和ML的方法。首先，它集成了一个姿势检测机器学习模型(PoseNet[8])，该模型由身体点(即左耳、鼻子等)组成。以及这些关键点的置信度分数，这些都可以在代码中访问。其次，DanceOn使学习者能够导入在谷歌的可教机器[2]中训练的模型，以访问概率分类器来操纵和触发动画。虽然我们已经探索了DanceOn的使用，使学习者能够构建与文化相关的艺术制品[5]，但我们还没有探索如何通过使用该系统来教授人工智能和ML的概念和伦理。我们通过与教师共同设计具有文化维持性的AI/ML资源来调查两个研究问题：RQ1)当学习者建立、探索和批评ML和AI模型和系统时，我们如何真正建立在学习者的身份、文化知识和舞蹈实践之上？RQ2)我们如何利用学习者在ML和AI方面的创造性、具体化体验，在他们的社区和更广泛的社会中促进对CS和AI/ML的反思和批评？我们把合作设计作为一种方法，通过在学习设计的实现、实施和评估中以教师的价值观、所有权和真实背景为中心来促进可持续性[7]。通过共同设计课程资源，我们将开发一套与AI4K12指导方针[1]相一致的模块，以建立我们对创建具有文化可持续性的教育资源的理解，以利用舞蹈的文化境遇、协作和具体化的本质来教授人工智能和伦理设计。
Artificial intelligence (AI) and machine learning (ML) systems are ubiquitous across many fields ranging from medicine (e.g., tumor detection), to natural language processing (e.g., digital home assistants, auto-translate tools), and personalization (e.g., social media recommendations). The rise of these systems has also seen a rise in the ethical challenges resulting from how some of these systems are designed and implemented. Most problematic is their propagation and amplification of problems such as racism and sexism [4, 6], among others. Further, these challenges exist within a computing discipline that is already burdened by exclusive, marginalizing cultures and practices that lead to low participation by women and Black, Indigenous, and People of Color (BIPOC) [3]. Our project, AI + Dance, attends to these dimensions of inequity within AI/ML education through developing our understanding of how we can equip learners to recognize and rectify issues of AI and ML systems within an inclusive and culturally sustaining experience. We explore this in collaboration with STEM From Dance1 (SFD), a non-profit organization that supports girls of color in creative production with dance, CS, and STEM. In our prior work with SFD, we developed danceON, an open-access creative coding environment that enables learners to create code to engage authentically with dance and body motion [5]. With danceON, learners can code animations that can bind and respond to body positions and be statically and dynamically positioned in space. The system provides two ways to explore AI and ML. First, it integrates a pose detection machine learning model (PoseNet [8]) that consists of body points (i.e., Left Ear, Nose, etc.) and confidence scores for these key points, which are all accessible within the code. Second, danceON enables learners to import a model trained in Google’s Teachable Machine [2] to access the probabilistic classifier to manipulate and trigger animations. While we have explored the use of danceON to enable learners to build culturally relevant artistic artifacts [5], we have yet to explore how we can teach AI and ML concepts and ethics through the use of the system. We investigate two research questions through the co-design of culturally sustaining AI/ML resources with teachers: RQ1) How can we authentically build on learners’ identities, cultural knowledge, and practices with dance as they build, explore, and critique ML and AI models and systems? RQ2) How can we leverage learners’ creative, embodied experiences with ML and AI to facilitate reflection and critique of CS and AI/ML within their communities and society more broadly? We focus on co-design as a method that facilitates sustainability through centering teachers’ values, ownership, and authentic contexts in the realization, implementation, and evaluation of learning designs [7]. Through co-designing curricular resources, we will develop a set of modules aligned with the AI4K12 guidelines [1] that builds our understanding of creating culturally sustaining educational resources to teach about AI and ethical design in ways that leverage the culturally situated, collaborative, and embodied nature of dance.
danceON：文化响应式创意计算
DOI：
10.1145/3411764.3445149
发表时间：
2021
期刊：
CHI '21: Proceedings of the 2021 CHI Conference on Human Factors in Computing Systems
影响因子：
--
作者：
Payne, William Christopher;Bergner, Yoav;West, Mary Etta;Charp, Carlie;Shapiro, R. Benjamin;Szafir, Danielle Albers;Taylor, Edd V.;DesPortes, Kayla
通讯作者：
DesPortes, Kayla