From Siri to Chat GPT, Artificial Intelligence (AI) is changing the way people plan their days, communicate with their friends and family, and more. But what to many might feel like a trend or a novelty is quickly becoming an essential tool and a solution for the public sector. In 2020, it was estimated that 150 federal government programs used AI to assist with decision making and to make predictions based upon vast amounts of data and algorithms. States and municipalities are also adopting AI in a variety of contexts including law enforcement, public benefits distribution, fraud detection, permitting, employment, housing, and more.
Local and state governments need solutions to address backlogs in public services for their residents, while also facing budget cuts at the state and local level. AI holds the potential to streamline government services—to be faster and more responsive to residents and constituents. Pairing quality data with responsible automation presents an efficient approach. Unfortunately, there are also numerous examples in which public sector algorithmic error has resulted in life-changing repercussions.
- Between 2013-15 Michigan state’s unemployment insurance agency falsely accused 40,000 people of unemployment insurance fraud because of an error-prone automated software system built by a private company.
- In California and Colorado in the mid-2010s, pregnant women and foster children were denied Medicaid based upon algorithms that used faulty data and hundreds of incorrectly encoded rules.
- It was reported in 2023 that the federal Internal Revenue Service was auditing Black taxpayers at three times the rate of other taxpayers based on systemic discrimination by an AI system.
Municipal, county, state, and federal governments are using AI to make critical decisions about their residents. However, the “blackbox” nature of some AI systems may limit one’s ability to provide accountability and ensure fairness.