Andrew Horton

By Andrew Horton

Causal economic machine learning (CEML) is a relatively new research stream of AI that sits at the intersection of innovative developments in causal machine learning (CML) and behavioral science. It explicitly aligns AI with human intelligence and builds in a consequential relationship between costs and benefits, both contemporaneously and over time, which represents a fresh lens on the ethics of AI. This brief chapter explores the ethical rationale and implications of CEML and provides practical guidance on ethical implementation with the objective of increasing the use of CEML in research. The ethical implications of CEML are primarily positive, including enhanced transparency and interpretability, improved fairness and accountability, and greater opportunities for human oversight to reduce risk. However, CEML also demands careful governance in practice, including thorough validation of causal assumptions, ongoing ethical scrutiny in order to avoid potential bias, and the misuse or overconfidence in AI-guided decisions. CEML has powerful applications in areas where outcomes are highly consequential and ethics play a prominent role, such as personalized healthcare and economic policy.

Part of the book: Artificial Intelligence Ethics, Law, and Policy [Working title]