The Joint Statistics Meetings had a great talk on Fairness in Artifical Intelligence (AI). The speaker, Sherrie Rose, documented instances where AI has had a discriminatory effect, the wide ranges of causes of this, and how to fix things. I looked up some of the references and resources that she mentioned in her talk and list them below.
The best place to start is Fairness, Accountability, and Transparency in Machine Learning, a group that organizes annual conferences on this topic. They have a very nice resource list. Also quite helpful is a working group within the American Statistical Association, the Justice, Equity, Diversity, and Inclusion (JEDI) Outreach Group.
Statistics Canada has a couple of valuable resources:
- From Exploring to Building Accurate Interpretable Machine Learning Models for Decision-Making: Think Simple, not Complex
- Responsible use of machine learning at Statistics Canada
Stories in the news media about this topic include
- David Grossman. Amazon Fired Its Resume-Reading AI for Sexism. Popular Mechanics, 2018-10-18. Available in html format.
- Sidney Fussell. Why Can’t This Soap Dispenser Identify Dark Skin? Gizmodo, 2017-08-17. Available in html format
- Kashmir Hill. Wrongfully Accused by an Algorithm. The New York Times 2020-06-04. Available in html format
There are various pre-prints and peer-reviewed publications that offer technical explanations of the problem and propose some solutions.
- Irene Y. Chen, Emma Pierson, Sherri Rose, Shalmali Joshi, Kadija Ferryman, Marzyeh Ghassemi. Ethical Machine Learning in Health Care. arxiv.org, 2020-10-08. Available in pdf format
- Darshali A. Vyas, Leo G. Eisenstein, David S. Jones. Hidden in Plain Sight <U+FFFD> Reconsidering the Use of Race Correction in Clinical Algorithms. New England Journal of Medicine, 2020-08-27, 383, 874-882, DOI: 10.1056/NEJMms2004740. Available in html format.
- Anna Zink, Sherri Rose. Fair Regression for Health Care Spending. arxiv.org, 2091-07-13. Available in pdf format
- Thomas McGuire, Anna Zink, and Sherri Rose. Improving the Performance of Risk Adjustment Systems Constrained Regressions, Reinsurance, and Variable Selection. American Journal of Health Economics, 2021-03-18.
- Savannah L.Bergquist, Timothy J.Layton, Thomas G.McGuire, SherriRose. Data transformations to improve the performance of health plan payment methods. Journal of Health Economics, 2019 (July), 66, 195-207.
- Caroline Wang, Bin Han, Bhrij Patel, Feroze Mohideen, Cynthia Rudin. In Pursuit of Interpretable, Fair and Accurate Machine Learning for Criminal Recidivism Prediction, arxiv.org, 2020-05-08. Available in pdf format