Puzzle-solving in AI Ethics: Theories, Principles, and Tools

Cansu Canca


May 11, 2022 / 1:00-2:00 p.m. EDT

Expeditions in Experiential AI

While AI ethics and responsible AI increasingly receive more attention, there remains a confusion about their scope and how to execute them. Multiple disciplines work on ethical questions related to AI but the disconnect between them impair the efficiency and robustness of the practice. A particularly interesting phenomenon in AI ethics practice has been the proliferation of AI ethics principles around the world. AI ethics principles can serve as a useful tool in helping ethical decision-making. But most published sets of AI ethics principles fail to be operational due to their conceptual errors and vagueness, and a lack of understanding of their function.

In this talk, I will situate AI ethics and its core questions within relevant disciplines. Zooming into AI ethics principles, I will show the central role of ethics literature and ethical theories in creating operable and guiding tools for AI ethics practice. I will also (briefly) argue that the practice of AI ethics resembles “puzzle-solving” and thus should be executed within a model of practice that allows critical, collaborative, continuous, and dynamic interaction of ethics and innovation rather than being limited to a compliance approach.



Cansu joined the Institute of Experiential AI as the ethics lead and a research associate professor. She also has an affiliation with the Department of Philosophy and Religion and the Ethics Institute in the College of Social Sciences and Humanities. She has a doctorate in philosophy specializing in applied ethics. 

Cansu is the founder of AI Ethics Lab, one of the first initiatives focusing exclusively on advising practitioners and conducting multidisciplinary research on the ethics of artificial intelligence. She remains the director of the AI Ethics Lab, where she leads teams of computer scientists, philosophers, legal scholars, and other experts in research, the development of toolkits, and consulting.

Cansu developed the Puzzle-solving in Ethics (PiE) Model, a dynamic and collaborative model for integrating ethics into the AI innovation cycle that organizations have implemented through consulting work with the AI Ethics Lab. She brings the PiE Model to EAI along with her industry experience and her background in philosophy.

Cansu serves as an ethics expert in various ethics, advisory, and editorial boards. She is a founding editor for the international peer-reviewed journal AI & Ethics (Springer Nature), serves as an ethics expert for EU-funded research projects focusing on the ethics of AI, robotics, human enhancement, and law enforcement AI technologies, and chairs the Institute of Electrical and Electronics Engineers (IEEE) AI Experts Network Criteria Committee.

Before her work in technology, Cansu spent more than a decade working in population-level bioethics on a range of topics including, resource allocation, human subject research, and end-of-life decisions. She was on the full-time faculty at the University of Hong Kong Medical School and a researcher at the Harvard Law School, Harvard School of Public Health, Harvard Medical School, National University of Singapore, Osaka University, and the World Health Organization. 

Frequently invited to speak to leaders in industry and academia, Cansu has given over 100 talks on AI ethics, including keynotes at Harvard Business School, the U.S. Department of Justice, her TEDxCambridge talk How to Solve AI’s Ethical Puzzles, and several others. She is among the 30 Influential Women Advancing AI in Boston and 100 Brilliant Women in AI Ethics.