Michigan Proposes Oversight Board for State AI Use
Michigan state government employees are using artificial intelligence without clear oversight, and a new bill aims to change that. House Bill 5899 would create a three-member governing board to study how generative AI is being used across state agencies and establish a pilot program by January 1, 2027.
State Rep. Jaime Greene, R-Richmond, said the current situation is "still kind of a mystery." Departments are using AI tools, but "nobody has really been revealing how they're utilizing AI right now," he said.
The bill passed the House Communications and Technology Committee unanimously on May 19 and has been referred to the House Rules Committee.
What the Pilot Program Would Do
The governing board would include one member with AI or data science expertise, one with a background in ethics or civil rights, and one from the private sector. The governor would appoint all three.
The board would help design a pilot program allowing state employees to test generative AI tools in controlled settings. Participants could not use the tools on restricted or confidential information.
State officials would measure time saved, efficiency gains, and any risks or consequences from AI use. The data would be reported back to the legislature.
Cost and Funding Questions
The House Fiscal Agency estimates the program would cost $600,000 to set up, with $2.1 million annually for staffing, software, licensing, and training.
Michigan lawmakers previously set aside $10 million for an information technology fund to develop AI platforms over four years. A Republican-controlled House committee cancelled that work project fund in December 2025, prompting a lawsuit. If a court permits the spending, those funds could support Greene's pilot program.
Greene said he is not seeking new appropriations. "This is asking DTMB to essentially do their job and study AI and its efficiency and efficacy in state government," he said.
How Other States Are Moving
Michigan is not alone. All 50 states, Puerto Rico, the Virgin Islands, and Washington, D.C. introduced AI legislation in 2025, according to the National Conference of State Legislatures.
Several states have already launched pilot programs. California created a digital AI assistant that saved employees "hundreds of hours," according to testimony from Kevin Frazier, a senior fellow at the Abundance Institute. New Jersey, North Carolina, North Dakota, Pennsylvania, and Vermont have similar initiatives.
An Ernst & Young survey found that 51% of federal, state, and local government employees used an AI application daily or several times a week as of September 2024.
What Michigan Has Done So Far
Michigan's state technology department has said employees can use AI to generate reports, but only if the information is already public. The department has emphasized protecting state data.
The state passed laws in 2023 requiring disclosure labels on AI-generated political campaign ads and banning AI-generated deepfake pornography without consent.
Meanwhile, communities across Michigan have pushed back against data center proposals. As of April, more than 1,500 square miles of the state are under moratoriums pausing data center consideration in 51 cities or townships.
For government professionals seeking to understand AI implementation, resources on AI for Government and Generative AI and LLM can provide context on best practices being adopted across states.
Your membership also unlocks:
