This project in the School of Computing Instruction hopes to leverage the campus as a living-learning environment and platform for transdisciplinary research and education: It aims to provide a holistic understanding of ethical considerations in AI by integrating ethics-oriented problem sets into Machine Learning class. This is in alignment with using the campus as a transdisciplinary platform where students from different disciplines can engage with ethical AI practices in a practical, hands-on manner, thereby enhancing the educational experience beyond the traditional settings of an AI class. The project hopes to enhance access, equity and inclusion on campus: It's approach to teaching ethical AI involves creating assignments that emphasize fairness, unbiased datasets, and the reduction of model bias. This educational method promotes equity by preparing students to develop AI models that avoid perpetuating existing social biases, thereby contributing to more equitable outcomes in AI application. The project's goals emphasize the development of assignments that prepare students to think critically about the ethical implications of AI, using sustainability and equity as core principles. This is achieved through the expansion of course coverage to include ethical AI principles, practical implementation of these principles in homework assignments, and fostering critical reflection among students about the societal impacts of AI. These goals resonate with the strategies of leveraging educational platforms for transdisciplinary research and enhancing equity and access, as they are designed to cultivate a generation of AI practitioners who are not only technically proficient but also ethically conscious and socially responsible.
The original plan for the project was focused on incorporating reflection questions to prompt students to think about the ethical implications of AI. However, as the project progressed, it became clear that technical questions were also necessary to provide a comprehensive understanding of the subject. This adjustment was made to ensure that students not only reflected on the ethical considerations but also applied these considerations technically in their work with AI systems. The incorporation of technical questions aimed to enhance students' ability to practically apply ethical principles in their AI and machine learning models. This change was vital in helping students to critically engage with the coursework and develop skills that are directly relevant to the challenges they will face in the field. By including both reflective and technical components, the course design became more robust and well-rounded. Furthermore, addressing SDG 3 (Good health and well-being), SDG 5 (Gender Equity), and SDG 10 (Reduced Inequalities) was added to the agenda.
The redesign process necessitated a deep consideration of the ethical implications of AI, which broadened the scope and depth of instruction. The effectiveness of these changes was assessed through a survey comparing the responses from the control group in Fall 2023 to those after the changes in Spring 2024. Questions in the survey assess students' understanding of the SDGs, their ability to handle ethical challenges in AI, the effectiveness of the curriculum in conveying sustainability concepts, and project efficacy. As showcased by the results, students are now more equipped to identify and mitigate biases in datasets and understand the importance of fairness in AI. The redesigned assignments seem to have allowed students to critically reflect on AI's societal implications, thus ingraining a sense of responsibility as future practitioners. The effect of the final project on student learning was not significant. To change that, we plan to propose the following questions to be put into perspective: (1) How does the project address the sustainability concepts (ethical considerations, addressing environmental impacts, and economic considerations)? (2) Did your solution contribute to a more sustainable response (e.g. reduce biases and promote ethical approaches in practice, address environmental impacts by introducing a more efficient way, or introduce a way that is more economically efficient)? Overall trends indicated increased familiarity with SDGs, more positive views on AI's sustainability impact, and a greater emphasis on ethical considerations in the curriculum from Fall 2023 to Spring 2024. The most significant change was in how homework assignments encouraged ethical thinking. The syllabus for Fall 2023 onwards will be modified to emphasize the importance of sustainability in student projects.
"The assessment of learning outcomes has evolved to include not just the technical proficiency in AI and machine learning but also an understanding of the ethical and societal impacts of these technologies. Progress is evaluated through tailored assignments, practical applications, and semester-wide projects that reflect these enhanced learning goals. It was also very fulfilling to observe the tangible improvement in students' understanding of ethical and sustainability aspects by comparing both control (Fall 2023) and active (Spring 2024) semester. In addition, in the lecture session dedicated to sustainability, it was fascinating to see students engage thoughtfully and critically in nuanced discussions about the ethical considerations and aftermaths of AI."