Algorithmic Decision-Making Safeguarded by Human Knowledge

TL;DR

This paper develops a framework for integrating human knowledge as a safeguard to improve algorithmic decisions, especially when algorithms face domain knowledge gaps, model errors, or data issues.

Contribution

It introduces a general analytical framework for augmenting algorithms with human insights to enhance decision quality in specific problematic scenarios.

Findings

Human guardrails are beneficial when algorithms lack domain knowledge.

Augmentation offers no advantage when algorithms are asymptotically optimal.

Human knowledge can mitigate issues like model misspecification and data contamination.

Abstract

Commercial AI solutions provide analysts and managers with data-driven business intelligence for a wide range of decisions, such as demand forecasting and pricing. However, human analysts may have their own insights and experiences about the decision-making that is at odds with the algorithmic recommendation. In view of such a conflict, we provide a general analytical framework to study the augmentation of algorithmic decisions with human knowledge: the analyst uses the knowledge to set a guardrail by which the algorithmic decision is clipped if the algorithmic output is out of bound, and seems unreasonable. We study the conditions under which the augmentation is beneficial relative to the raw algorithmic decision. We show that when the algorithmic decision is asymptotically optimal with large data, the non-data-driven human guardrail usually provides no benefit. However, we point out three common pitfalls of the algorithmic decision: (1) lack of domain knowledge, such as the market competition, (2) model misspecification, and (3) data contamination. In these cases, even with sufficient data, the augmentation from human knowledge can still improve the performance of the algorithmic decision.