Co-developed in Philosophy and the School of Computing, a new class addresses today’s AI issues.
Embedded ethiCS and ‘agressive interdisciplinarity’ The University of Utah's new AI Ethics class draws inspiration from Harvard's 'Embedded EthiCS' initiative, a model born from student demand to address the moral outcomes of technology. Professor C. Thi Nguyen characterizes this approach as 'aggressive interdisciplinarity,' advocating for philosophers and computer scientists to collaborate in developing and teaching course materials. This ensures that ethical considerations are integrated into the fundamental process of writing code, rather than being an afterthought. Nguyen highlights that while philosophers identify abstract ethical problems, computer science faculty provide the practical engineering perspective to devise actionable solutions. The cross-listed class (PHIL 2150 / CS 2395) will be co-taught by faculty from both Philosophy and the Kahlert School of Computing, starting in Fall 2026. It is designed to attract a diverse student body, with a mix of computer science majors and students from other disciplines like philosophy, psychology, and communications. Students were already asking these questions Professor Nguyen notes a significant shift in student inquiries, with increasing concerns about AI's impact on moral and social skills, inherent biases in systems, and job displacement. To illustrate these ethical dilemmas, Nguyen uses real-world case studies. One example involves a machine learning system designed to predict student success, which was found to be biased by defining success too narrowly (graduation rate and speed) due to limited input from only deans. Another case is the COMPAS software, a risk assessment tool in the U.S. criminal justice system that produced racially biased recidivism predictions, demonstrating how existing societal biases can be reflected in data-driven systems, despite non-racist intentions. Nguyen refers to this as 'value-ladenness,' where technologies, seemingly neutral, make politically charged decisions based on their hidden design. The class also addresses 'moral and social deskilling,' where over-reliance on automation can diminish human capabilities. The use of ChatGPT as a therapist is cited as an example, as the AI's 'sycophantic' nature could prevent users from receiving necessary criticism, which Nguyen argues is essential for improvement and understanding diverse perspectives—a concern he finds more pressing than hypothetical 'killer AI'. Inspiring goodness Professor Nguyen views the widespread media focus on 'killer, world-ending AI' as a potential 'soft marketing campaign' that diverts attention from more immediate and significant concerns such as job displacement, AI's environmental impact (like water consumption), and the erosion of social life. He suggests that this narrative inadvertently benefits Silicon Valley by emphasizing AI's perceived power. Despite these media-driven narratives, Nguyen maintains optimism about the future of artificial intelligence and its development, particularly through the education of his students. He acknowledges that the definition of 'good' for AI can vary significantly across cultures and nations, making international regulations crucial for managing risks. However, he emphasizes that embedding ethics within the code through education is a more proactive strategy. While regulations can prevent negative outcomes, a classroom setting can actively 'inspire people to be aggressively good and sensitive' in their development of burgeoning technologies, fostering a generation of engineers who prioritize ethical design from the ground up.
Embedded ethiCS and ‘agressive interdisciplinarity’
The University of Utah's new AI Ethics class draws inspiration from Harvard's 'Embedded EthiCS' initiative, a model born from student demand to address the moral outcomes of technology. Professor C. Thi Nguyen characterizes this approach as 'aggressive interdisciplinarity,' advocating for philosophers and computer scientists to collaborate in developing and teaching course materials. This ensures that ethical considerations are integrated into the fundamental process of writing code, rather than being an afterthought. Nguyen highlights that while philosophers identify abstract ethical problems, computer science faculty provide the practical engineering perspective to devise actionable solutions. The cross-listed class (PHIL 2150 / CS 2395) will be co-taught by faculty from both Philosophy and the Kahlert School of Computing, starting in Fall 2026. It is designed to attract a diverse student body, with a mix of computer science majors and students from other disciplines like philosophy, psychology, and communications.
Students were already asking these questions
Professor Nguyen notes a significant shift in student inquiries, with increasing concerns about AI's impact on moral and social skills, inherent biases in systems, and job displacement. To illustrate these ethical dilemmas, Nguyen uses real-world case studies. One example involves a machine learning system designed to predict student success, which was found to be biased by defining success too narrowly (graduation rate and speed) due to limited input from only deans. Another case is the COMPAS software, a risk assessment tool in the U.S. criminal justice system that produced racially biased recidivism predictions, demonstrating how existing societal biases can be reflected in data-driven systems, despite non-racist intentions. Nguyen refers to this as 'value-ladenness,' where technologies, seemingly neutral, make politically charged decisions based on their hidden design. The class also addresses 'moral and social deskilling,' where over-reliance on automation can diminish human capabilities. The use of ChatGPT as a therapist is cited as an example, as the AI's 'sycophantic' nature could prevent users from receiving necessary criticism, which Nguyen argues is essential for improvement and understanding diverse perspectives—a concern he finds more pressing than hypothetical 'killer AI'.
Inspiring goodness
Professor Nguyen views the widespread media focus on 'killer, world-ending AI' as a potential 'soft marketing campaign' that diverts attention from more immediate and significant concerns such as job displacement, AI's environmental impact (like water consumption), and the erosion of social life. He suggests that this narrative inadvertently benefits Silicon Valley by emphasizing AI's perceived power. Despite these media-driven narratives, Nguyen maintains optimism about the future of artificial intelligence and its development, particularly through the education of his students. He acknowledges that the definition of 'good' for AI can vary significantly across cultures and nations, making international regulations crucial for managing risks. However, he emphasizes that embedding ethics within the code through education is a more proactive strategy. While regulations can prevent negative outcomes, a classroom setting can actively 'inspire people to be aggressively good and sensitive' in their development of burgeoning technologies, fostering a generation of engineers who prioritize ethical design from the ground up.
