Michigan: building AI literacy across public schools through state-endorsed guidance and teacher training

The problem

Students currently in school will enter a job market where AI is a routine part of how organisations operate. Whether they go into healthcare, public administration, manufacturing, or education, they will encounter AI in their working lives. Understanding what it is, what it can and cannot do, and how to use it responsibly is becoming as fundamental as digital literacy was a generation ago.

For any of that to reach students, teachers need to understand AI themselves, and most have had little chance to learn it. The state also faced a consistency problem: with many AI resources on the market, each district was left to judge for itself what to trust, which risked uneven quality across the system.

Michigan faced a practical question: how do you prepare an entire state's teaching workforce for something that most of them were never trained in? A small team of specialists could not cover every school and every district. Sending every teacher on an individual training course would be slow and expensive. The state needed an approach that could efficiently build knowledge across the system, reaching educators in every region regardless of how well-resourced their schools were.

The solution

Michigan's approach has two parts: the state designated a single, trusted set of guidance, and it relied on a state-funded body to build teachers' skills.

First, the Michigan Department of Education endorsed Michigan Virtual's AI resources as the official AI guidance for its school districts. That gave every district a recognised reference point to work from, rather than assessing competing materials on its own.

Second, the work is led by Michigan Virtual, a non-profit that the state funds and that reports to the Michigan Legislature and the Department of Education, through its AI Lab. To extend that work to more teachers, Michigan Virtual partnered with the AI Education Project (aiEDU), a national non-profit, on a train-the-trainer model: instead of teaching every educator centrally, prepare a smaller group who can train their colleagues.

Training trainers, not just teachers

The core of the partnership is a year-long programme running from June 2025 to June 2026. Fifty Michigan educators are selected and supported to become AI teaching leaders for their local areas. They work in two virtual learning cohorts, building practical skills and strategies for introducing AI concepts to students, then deliver foundational AI literacy sessions to colleagues in their own schools, so the knowledge spreads outward from 50 people to the educators around them. Participants earn State Continuing Education Clock Hours, so the training counts towards their existing professional development requirements.

Resources designed for every subject

The materials are not aimed only at computer science teachers. AI raises questions about ethics, fairness, privacy, and decision-making that are relevant across the curriculum. Michigan Virtual and aiEDU provide free, openly accessible curriculum for teachers in any subject area, including a Student Guide to AI, a Teacher Guide, and an Administrator Guide for school leaders. They are free to access, so schools do not need to secure funding before they can start.

The impact

The reach below comes from Michigan Virtual's broader AI work, led by its AI Lab, which predates and extends beyond the aiEDU partnership. The Train-the-Trainer programme began in mid-2025 and ran to June 2026, so its own outcomes are not yet available.

1. Over 19,000 educators reached across the state

Michigan Virtual reports that its AI Lab has delivered more than 500 workshops, keynotes, and professional development sessions, reaching over 19,000 educators.

2. The state has one official reference point for AI in schools

By endorsing Michigan Virtual's resources as its official AI guidance, the Department of Education gave all districts a single, recognised basis for approaching AI, rather than each district judging competing materials on its own. Without this, individual districts would each need to assess which of the many available AI resources to trust, leading to uneven quality and inconsistent approaches across the state.

3. A recurring statewide gathering on AI in schools

The 3rd Annual AI Summit in October 2025 drew over 500 educators from across the state, the third year of an annual event that brings teachers, administrators, and technology leaders together around AI in K-12 education.

Lessons learned

• Endorsing one trusted set of guidance reduces fragmentation. By naming a single recognised reference point, the state spared every district the task of judging competing AI materials on its own. For governments worried about uneven or unreliable AI resources reaching schools, designating trusted guidance is a low-cost lever.

• Training trainers is more scalable than training every teacher individually. The programme invests in 50 educators who then train colleagues across their own districts. For a government building a new capability across a large public workforce, this multiplier model reaches further than a central team can alone.

• Free and open resources support equity. Because the materials are openly accessible at no cost, less well-resourced districts can start at the same time as everyone else. In a system where school funding varies widely, who benefits first depends on that.

Launch year: 2023

This case study was written with assistance from artificial intelligence.