Oak Park’s trustees are right to pause before adopting an AI chatbot as part of the village’s renewed Granicus contract. The proposed “Government Experience Agent” is being sold as a way to provide accurate answers to residents while saving staff time. That promise sounds appealing, but it rests on assumptions that have not been proven in municipal settings.

Other cities offer cautionary lessons. New York City scaled back several automated response tools after residents reported confusing or incomplete answers on housing and benefits. Los Angeles found that chatbots worked only for narrow tasks like office hours, while complex service questions still required staff intervention. Boston and several UK local councils have similarly treated public-facing AI chat tools as limited pilots rather than core service infrastructure, citing ongoing maintenance demands and quality-control concerns.

Even the village’s own memo concedes that errors occur when source information is incomplete or outdated — a common condition for government websites. In that scenario, a chatbot does not invent information, but it can still deliver confident, incomplete, or misleading answers. For residents navigating taxes, permits, fines, or elections, that is not a minor risk.

There is also the cost. With onboarding fees, annual subscriptions, and a contract extending through 2031, Oak Park would be committing to a long-term vendor relationship without clear benchmarks for success or a clearly defined problem the tool is meant to solve.

Technology can improve government services — but only when adopted deliberately, not because “the world’s changing.” Being cautious here is not resisting progress; it is protecting residents from becoming test cases for an unproven solution.

Works cited:

City of New York. “Automated Decision Systems Task Force Final Report.” Mayor’s Office of Data Analytics, 2019.

Katz, Josh, and Matthew Haag. “When Algorithms Decide Who Gets Housing, New York Asks: Who Decides the Algorithm?” The New York Times, 21 Nov. 2019.

City of Los Angeles Information Technology Agency. “Customer Service and Digital Tools Assessment.” City of Los Angeles, 2021.

City of Boston Analytics. “Responsible Artificial Intelligence: Municipal Use Cases and Limitations.” City of Boston, 2022.

UK National Audit Office. “Challenges in Using Artificial Intelligence in Public Services.” National Audit Office, 2020.

Bob Milstein
Oak Park

Join the discussion on social media!