Get out of the BAG! Silos in AI Ethics Education: Unsupervised Topic Modeling Analysis of Global AI Curricula

Abstract

The domain of Artificial Intelligence (AI) ethics is not new, with discussions going back at least 40 years. Teaching the principles and requirements of ethical AI to students is considered an essential part of this domain, with an increasing number of technical AI courses taught at several higher-education institutions around the globe including content related to ethics. By using Latent Dirichlet Allocation (LDA), a generative probabilistic topic model, this study uncovers topics in teaching ethics in AI courses and their trends related to where the courses are taught, by whom, and at what level of cognitive complexity and specificity according to Bloom’s taxonomy. In this exploratory study based on unsupervised machine learning, we analyzed a total of 166 courses: 116 from North American universities, 11 from Asia, 36 from Europe, and 10 from other regions. Based on this analysis, we were able to synthesize a model of teaching approaches, which we call BAG (Build, Assess, and Govern), that combines specific cognitive levels, course content topics, and disciplines affiliated with the department(s) in charge of the course. We critically assess the implications of this teaching paradigm and provide suggestions about how to move away from these practices. We challenge teaching practitioners and program coordinators to reflect on their usual procedures so that they may expand their methodology beyond the confines of stereotypical thought and traditional biases regarding what disciplines should teach and how.
 This article appears in the AI & Society track.

Ethical issues of the use of AI-driven mobile apps for education

“…The chosen studies can be further subdivided into those that address purely educational issues related to AI ( 19 , 21 , 22 , 24 ) and those that discuss the use of AI in medicine ( 20 , 23 ) and could also be included in this study, while Leimanis and Palkova ( 18 ) address ethical issues of AI in medical education. The remaining study ( 14 ) deals with the ethics of AI in various disciplines, with machine ethics, a subfield of the ethics of AI, as a focal point but it provides insightful comments related to education, therefore, it also had to be included in this review.…”

“…The most widespread ethical theories are mentioned in at least one of the examined texts. These include deontology ( 14 , 20 , 21 ), utilitarianism [( 14 , 20 , 21 ), consequentialism ( 14 ), virtue ethics ( 14 , 21 )]. These conceptual clarifications seem relevant and provide the reader with a necessary theoretical background that can be further extended into a practical application of this theoretical framing.…”

“…Javed et al ( 21 ) studied the disciplinary distribution of AI ethics course delivery by the department and argue that AI courses are mainly delivered by computer science and humanities departments, but also by law departments, and some courses are multidisciplinary, i.e., offered by at least two different departments. The same authors recognize three basic approaches to teaching AI ethics, namely Build (engineering and technical aspects—trustworthy technical solutions), Assess (multidisciplinary teams focused on philosophy and application—judgments based on fundamental principles), and Govern (humanities and law disciplines focused on developing general knowledge—stakeholder protection).…”

“…The same authors recognize three basic approaches to teaching AI ethics, namely Build (engineering and technical aspects—trustworthy technical solutions), Assess (multidisciplinary teams focused on philosophy and application—judgments based on fundamental principles), and Govern (humanities and law disciplines focused on developing general knowledge—stakeholder protection). They add that holistic teaching that removes disciplinary, topic-oriented, and other barriers should be pursued ( 21 ). In the same vein, Stenseke ( 14 ) emphasizes the importance of avoiding a narrow disciplinary perspective by analytically understanding the different ways in which each discipline understands underlying concepts, conducts its research, and produces results.…”

“…While all stakeholders can indicate various practical ethical problems related to AI and offer possible solutions, ethicists are there to refine all of this with respect to ethical theory and to point out possible pitfalls of a lay perspective. Still, even if there are (globally accepted) ethical guidelines for AI [e.g., Javed et al ( 21 ) mention the ACM ethics guidelines—see https://ethics.acm.org/code-of-ethics/software-engineering-code/ ], they will not necessarily lead to the ethical functioning of AI. Businesses and institutions can abuse or misuse them as ethics-washing, i.e., a strategy to cover up unethical behavior ( 14 ).…”