The Ethics of Biased Artificial Intelligence: A Stakeholder and Shareholder Theory Investigation

Loading...
Thumbnail Image
Penn collection
The Wharton School::Wharton Undergraduate Research::Wharton Research Scholars
Degree type
Discipline
Business
Subject
Ethics
Artificial Intelligence
Funder
Grant number
Copyright date
2022
Distributor
Related resources
Author
Ria Chinchankar
Contributor
Abstract

This paper aims to investigate how shareholder and stakeholder theory apply to issues of biased artificial intelligence, specifically machine learning. It is done through an event study that measures whether a significant abnormal return in share price occurs on the day of a news leak about a biased AI event, which can be used to understand how market discipline and thus shareholder theory apply in this context. A sample of ten large publicly listed US-based technology companies was used. No consistent, statistically significant abnormal returns were found, which indicates no evidence that shareholder theory is an effective mechanism for preventing and responding to issues of biased artificial intelligence. This troubling finding prompts further exploration of stakeholder theory and government regulation as methods to ensure justice and fairness in algorithmic decision making.

Advisor
Date Range for Data Collection (Start Date)
Date Range for Data Collection (End Date)
Digital Object Identifier
Series name and number
Publication date
2022-05-01
Volume number
Issue number
Publisher
Publisher DOI
Journal Issue
Comments
Recommended citation
Collection