Automated Decision-making

Automated decision-making systems use algorithms, artificial intelligence, and machine learning to make or assist with decisions that traditionally required human judgment. Governments and private companies are increasingly adopting these systems to determine eligibility for public benefits, assess credit and insurance applications, screen job candidates, set prices, and allocate resources. While these systems can improve efficiency, they also raise concerns about bias, lack of transparency, and potential discrimination, especially in critical areas like housing, employment, and criminal justice.

Regulation and Oversight

Several states and cities, including Colorado, Illinois, and New York City, have passed laws regulating automated decision-making to prevent algorithmic discrimination. These laws often require impact assessments, transparency, and protections against unfair outcomes. Federal agencies enforce existing laws targeting biased algorithms in housing and employment. A notable example is the DOJ settlement with RealPage, which addressed concerns about rental pricing algorithms affecting competition and fairness in housing markets.

Government Use Cases

Automated decision-making is also used by public agencies for fraud detection, resource allocation, eligibility determinations for benefits, and risk assessments in social services. Many government offices use these systems embedded in standard software, sometimes without full awareness of their implications. To address this, states are implementing audits, inventories of automated systems, and procurement guidelines to improve accountability. Balancing efficiency with transparency and fairness remains a key challenge in deploying these technologies.