How Michigan’s New Auto-Scheduling Push Could Reshape Due Process for High-Volume Dockets

Michigan is automating court scheduling for high-volume dockets. The public record says almost nothing about override authority, exception handling, or what happens to litigants when the system gets it wrong.

What the 2026 JIS Report Actually Says

The Michigan Judicial Information Services Legislative Report for 2026 is a systems and progress document. It tells the legislature where the statewide case management system stands, what was completed in the prior year, and what is planned going forward. On the question of scheduling automation, its language is straightforward: for high-volume dockets, JIS is enhancing the CMS to allow certain court activities and hearing types to be auto-scheduled following court rules, policies, and judge availability, complemented by electronic notification.^[1]

That is the entirety of the public description. It does not say which docket types will be auto-scheduled first. It does not describe the rule set the system applies when calendaring a hearing. It does not explain what triggers an exception, who can override an automated date, or what the litigant-facing process looks like when automation produces a date that creates a genuine conflict or a notice failure.

Between March 2025 and February 2026, ten courts migrated to the statewide CMS. As of February 2026, fifteen more are in active implementation, and another ten are in early discussions with JIS.^[2] In 2026, more than sixty probate courts are scheduled to move from legacy systems to the web-based CMS, making them eligible for MiFILE integration for the first time.^[3] These are not projections. They are the current rollout pace. And the governance structures that would define how automated scheduling interacts with due process are still being stood up: the Judicial Information Advisory Council, chartered to govern statewide CMS implementation, convened for the first time in the first quarter of 2026.^[4]

Governance is being built while the system is already running. That sequencing matters.

Automation Is Not Neutral When Court Access Is Unequal

Court scheduling is not a neutral administrative function. A hearing date is a legal event with consequences attached: the obligation to appear, the right to notice, the opportunity to be heard. When scheduling is manual, there is a human in the chain who can observe a conflict, flag a problem, or catch an error before it becomes a missed appearance. When scheduling is automated, the error-catching depends on the quality of the rules, the robustness of the exception logic, and the existence of someone whose job it is to monitor for failure.

Michigan’s rules on e-filing create an immediate asymmetry that automated scheduling inherits. Attorneys practicing in MiFILE courts are mandated to e-file under MCR 1.109(G)(3)(f). Self-represented litigants may e-file but are not required to do so. Courts must make public workstations available for in-person access to electronic records.^[5] When automated scheduling generates a hearing date and dispatches electronic notification, parties who are not digitally engaged with the system may receive notice through a different channel, on a different timeline, or not at all if their contact information is not current in the CMS.

This is not a hypothetical risk. It is a documented failure mode in systems that combine digital-first notification with populations who are not consistently reachable digitally. The populations most likely to fall into that gap in Michigan’s courts are people navigating family law matters, traffic and landlord-tenant cases, and probate proceedings, precisely the docket types the JIS report identifies as candidates for early automation due to high volume.

Where Michigan’s Own Research Points

Michigan is not building automated court systems in an evidence vacuum. The Michigan Judicial Council convened a Generative AI and the Courts Workgroup in 2024, which issued a detailed report before auto-scheduling was confirmed as a CMS feature. That report is specific about the risks that attach to algorithmic systems in judicial settings.^[6]

On bias, the report identifies that AI models can perpetuate or amplify existing biases if not carefully monitored, and that this risk is particularly acute in areas where automated outputs influence judicial decisions affecting individual rights. On transparency, the report flags the black-box problem directly: when an algorithm’s decision logic cannot be explained in terms a litigant can understand and contest, it may conflict with the principle of open justice and the right to a fair trial. On pro se litigants specifically, the report notes that AI-related deficiencies appear more frequently in filings submitted by parties without counsel, a pattern that would transfer to automated notice failures if the underlying system cannot distinguish between a party who did not receive notice and a party who chose to ignore it.

These are not abstract concerns from an outside critic. They are the findings of a workgroup convened by the same institution building the system. The question is whether the automated scheduling implementation actually incorporates those findings, or whether the risk analysis and the CMS roadmap are running on parallel tracks.

Where Is the Human Safety Valve?

Every scheduling system, manual or automated, produces conflicts. A witness becomes unavailable. A defense attorney has a scheduling conflict across two courts. A litigant receives a date they cannot make because of work, childcare, or medical obligations. In a manually scheduled system, the answer to those conflicts involves a motion, a clerk, a judge, and a human decision about whether to accommodate the request. The process is slow and imperfect, but it has a legible decision-maker and a reviewable record.

In an automated system, the equivalent safety valve is the override procedure. Who can initiate one. What the standard is. How quickly it operates. Whether the litigant has meaningful notice that an automated date has been set before the window to contest it closes. None of this is described in the 2026 JIS report. None of it appears in the publicly available MiFILE briefs or implementation documentation reviewed for this analysis.

The 12 FTE positions SCAO requested for FY 2026 to support the statewide CMS expansion were denied.^[7] Those positions were described as essential to support, update, enhance, and secure the system long-term. The positions SCAO sought included roles specifically tied to system governance and court support. Their absence does not mean automated scheduling is unsupported. It does mean the human oversight layer is thinner than the agency responsible for the system believes it should be.

Questions Michigan Courts Should Answer Now

Efficiency is a legitimate goal. High-volume dockets are genuinely difficult to manage, and technology that reduces scheduling burden for clerks and judges has real value. The question is not whether to automate. The question is what accountability structures exist around automation before it becomes the operating default.

At minimum, the public record should contain answers to the following: Which docket types and hearing categories are subject to automated scheduling, and under what court rules? What is the override process when a litigant or attorney has a legitimate conflict with an automatically generated date, and who has authority to exercise it? How does the system ensure that self-represented litigants who are not registered in MiFILE receive equivalent notice to parties who are? What happens when an automated notice is not received, and who has standing to raise the failure after the fact? What monitoring exists to detect bias, error rates, or disparate outcomes across case types and litigant populations?

These are not adversarial questions about technology. They are the minimum transparency requirements for any process that generates legal obligations. The fact that they are not answered in the current public record is the problem, not the automation itself.