Principles for Fairer, More Responsive Automated Decision-Making Systems

Executive Summary

Many government agencies use automated decision-making systems (ADS) for public benefits coverage and enrollment decisions, including Medicaid. By ADS we mean, broadly, systems that use standardized logic or algorithms to assess eligibility for and access to public benefits. This includes not only sophisticated “big data” systems, such as programs that use machine learning to flag potentially fraudulent applications, but also much simpler protocols that enable coverage decisions with little or no active human judgment, such as evaluating eligibility criteria using available data.

NHeLP’s long history of advocacy has encouraged us to think about preventive advocacy rather than only addressing ADS after they have begun to harm individuals.

The broad principles described in this issue brief outline how to realize the benefits of ADS while minimizing drawbacks. From ensuring meaningful beneficiary input during design to incorporating transparency, effective due process protections, privacy protections, beneficiary outreach, and system oversight after implementation, the principles cover the complexities of building successful ADS.

Related Content