Professional Judgment in an Era of Artificial Intelligence and Machine Learning

Abstract

There are two fundamental features of the information processing behind most efforts to substitute artificial intelligence, machine learning, and robotics for professionals in health and education: reductionism and functionalism. However, true professional judgment hinges on a way of knowing the world and relating to persons that is at odds with the mindset of substitutive automation. Instead of reductionism, an encompassing holism is a hallmark of professional practice — an ability to integrate facts and values, to respect the demands of the particular case and prerogatives of society, and to balance mission and margin in institutional decision-making. Any presently plausible vision of substituting artificial intelligence for education and health care professionals would be premised on patients and students accepting services as “medical care” or “education” that are often far inferior to what a skilled, reflective practitioner in either field could provide. The only way these sectors can progress is to maintain, at their core, a large (and likely growing) core of professionals capable of carefully intermediating between technology and the patients it would help treat, or the students it would help learn. As critical data studies have repeatedly shown, the lifeblood of AI ambitions — data — is neither brute nor given. Deciding what data matters, how it is fairly and accurately collected, and how to balance quantitative and qualitative approaches to the representation of situations, will be critical and enduring roles for professionals.

Keywords: Automation, Epistemology of Automation, Taylorism, Disruptive AI, Algorithmic Decision Making, Clinical Practice Guidelines, Computational Thinking, Rankings, Metrics, Data-Driven