LANSING, Mich. (WZMQ) – The Michigan House Communications and Technology Committee heard testimony Tuesday on legislation that would create a statewide artificial intelligence pilot program aimed at studying how generative AI could be used across state government.
The bill, introduced by Rep. Jaime Greene (R-Richmond), would require the Department of Technology, Management and Budget, or DTMB, to establish a pilot program allowing state employees, agencies, and departments to test approved AI systems under state oversight.
Supporters said the proposal is designed to improve efficiency while also creating safeguards and transparency around how AI is used inside government operations.
“Wouldn’t it be wonderful to actually have a legitimate report that showed if there were workflow and efficiency improvements, improvements of public service, improvements of customer service,” Greene told the committee.
The bill would create a three-member Artificial Intelligence Governing Board appointed by the governor. The board would include experts in artificial intelligence or data science, ethics or privacy, and the private sector. The board would advise DTMB on the design and implementation of the program and recommend ethical guidelines for AI use in state government.
Under the legislation, state employees participating in the pilot program could not input restricted or confidential information into AI systems unless the systems meet state privacy and cybersecurity standards. Human review would also be required before AI-generated outputs could be used.
Greene said lawmakers currently have limited information about how state agencies may already be using AI technology internally.
During committee questioning, Rep. Noah Arbit asked whether departments had confirmed if they are currently using AI tools.
“Correct,” Greene responded after saying departments had not clearly disclosed their AI use practices.
The bill would require DTMB to submit a public report within 180 days after the pilot program concludes. The report would include information on time saved, efficiency improvements, risks, and unintended consequences discovered during the testing process.
A House Fiscal Agency analysis estimated the pilot program could cost more than $2.1 million annually depending on the scope of implementation, staffing, software subscriptions, and AI systems selected for testing. Additional one-time startup costs, including training and software tracking systems, were estimated at about $600,000.
The committee also heard testimony from Kevin Frazier of the Abundance Institute, who said other states have already seen measurable benefits from AI pilot programs.
Frazier pointed to programs in states including North Carolina, Pennsylvania, Vermont, and California, where AI has been used to speed up permit reviews, assist employees with research tasks, and improve internal government operations.
“These tools can drastically improve state operations in a transparent, reliable, and significant fashion,” Frazier said. “Michigan ought to follow the lead of its sister states.”
Frazier also recommended possible revisions to the legislation, including expanding the governing board beyond three members and clarifying how human review requirements would work for more advanced AI systems.
If approved, the pilot program would begin by Jan. 1, 2027, and continue until DTMB determines the program’s goals have been fulfilled. The committee did not vote on the bill Tuesday.
