Michigan’s Safe Prisons Initiative is not the wrong idea. It is an incomplete one deployed into a system that needed a structural overhaul, not a press release. Assault numbers are up. Staff vacancy rates run between 15 and 35 percent at half the state’s facilities. Correctional officers are working mandatory 16-hour shifts and leaving faster than replacements can be trained. The department’s response is package scanners, mail photocopying, and technology upgrades. The officers who actually work in these facilities are telling anyone who will listen that scanners do not break up fights, don’t fill overnight shifts, and don’t stop the burnout feedback loop that is hollowing out institutional knowledge one resignation at a time. And underneath all of it is a question no one in Lansing is publicly ready to ask: whether a significant portion of the people filling these facilities need to be there at all.
The Rollout and the Immediate Blowback
On March 20, 2026, the Michigan Department of Corrections announced its Safe Prisons Initiative, framing it as a data-driven, multifaceted response to a documented safety problem. The data is not in dispute: assaults on employees climbed from 299 in 2024 to 355 in 2025, a nearly 19% increase. Assaults on prisoners rose from 481 to 527 in the same period. The department cited this data, correctly, as requiring a response.
The initiative’s five focus areas cover prisoner classification and bed space, contraband technology, prisoner programming, training, and recruitment and retention. MDOC Director Heidi Washington described it as the department’s “best” effort on behalf of staff and incarcerated people alike. Deputy Director Jeremy Bush said the department had been analyzing data on trends and incidents and would continue evaluating effectiveness.
The pushback from officers was immediate and specific. Corrections1, the corrections industry publication, surveyed readers directly: what would it actually take to recruit and retain people in this job? The responses did not mention scanners. They cited pension elimination, mandatory overtime that has officers working consecutive 16-hour shifts, pay that fails to compete with comparable public employment, and leadership that, in the words of multiple respondents, prioritizes programming for incarcerated individuals over basic respect for the staff who keep the facilities operating.
Byron Osborn, president of the Michigan Corrections Organization, took the more measured line of being “cautiously optimistic,” while specifically directing attention to prisoner classification as the operational issue requiring the most urgent attention. The union’s documented position, going back through multiple legislative cycles, is that dangerous inmates have repeatedly been placed in lower-security settings where the staff-to-inmate ratio and facility design cannot adequately manage the risk level they actually present. That is a population management problem, not a contraband problem.
What the Research Actually Says About Staffing and Safety
The operational logic that corrections officers are articulating is not anecdotal. It has a research foundation that the department’s initiative does not appear to have engaged with directly.
The National Institute of Justice’s synthesis of correctional officer safety and wellness research documented the mechanism clearly: diminished work performance, burnout, and absenteeism produce higher incarcerated-person-to-officer ratios and reduced security levels across entire facilities. This is not a staffing inconvenience. It is a security degradation pathway. Officers cited in NIJ-supported research described being asked to accomplish more with fewer resources, with predictable consequences for their mental health and their ability to do the job safely.
A 2025 systematic review published in Trauma, Violence, and Abuse examined 21 studies on correctional officer victimization and found that facility-level factors, including programming availability and security level configuration, are directly associated with assault rates on staff. A separate NIJ-supported literature review found that staffing shortages and high turnover create cycles in which low officer-to-inmate ratios and loss of experienced personnel threaten the facility’s ability to implement security requirements. A federal OIG report covering the Bureau of Prisons identified understaffing as a direct contributor to rising inmate violence, noting specifically that reduced morale and staff attentiveness resulting from mandatory overtime “decreased the overall safety of the institution.” Michigan’s situation mirrors this pattern at scale.
The piece that does not show up in contraband data is institutional knowledge. An experienced corrections officer knows which individuals in a unit are in conflict. They know the behavioral signals that precede incidents. They know the relationship dynamics that, if managed early, keep a situation from escalating. When an officer with 12 years of experience leaves and is replaced, or not replaced, by someone who just completed an eight-week training academy, that knowledge does not transfer. Byron Osborn himself noted that the department has become increasingly reliant on chemical agents and tasers as first-response tools, partly because the hands-on de-escalation skills that come with experience are less present in a rotating, overtime-exhausted workforce.
The Safe Prisons Initiative assumes the primary variable driving safety outcomes is contraband flow. The research on prison safety says the primary variable is the human system: officer-to-inmate ratios, staff experience levels, and facility stability. Technology can reduce one category of risk. It cannot compensate for a workforce that is burned out, under-experienced, and leaving faster than replacements can be trained. When those two assumptions collide in an understaffed facility at 11 PM on a mandatory double shift, the scanner is not what stands between order and a violent incident.
The Policy Overhead Problem: New Mandates, Same Depleted Workforce
There is a compounding problem embedded in the initiative’s contraband-reduction components that has received almost no coverage.
Every new policy creates operational overhead. The January 2026 legal mail photocopy policy requires mail handlers to open each piece of correspondence in front of the recipient, create a photocopy, and shred the original on the spot. TextBehind DOCS requires staff to scan QR codes on privileged mail to verify sender identity before delivery. Package scanning requires trained personnel to operate equipment and document results. These are not passive systems. They require time, attention, and personnel to execute correctly.
Michigan has added mail processing requirements, scanning protocols, and documentation layers to facilities that are already operating with vacancy rates between 15 and 35 percent. The officers expected to implement these new procedures are the same officers working mandatory overtime, managing reduced yard time because there are not enough staff to safely run activities, and handling unit supervision at ratios that the research literature identifies as directly dangerous. Adding process requirements to a depleted workforce does not improve safety. It redistributes the same limited capacity across a larger set of obligations, leaving less of it for direct supervision and early conflict intervention, which is where the actual safety value lives.
This is a classic systems failure pattern: a visible, measurable problem gets a visible, deployable solution, while the less visible but more systemic problem continues to compound in the background. Contraband can be photographed and quantified. Burnout cannot be photographed. A resignation spike does not generate a press release. So the system defaults to what it can announce and measure, and the harder problem keeps getting worse.
The Population Question Nobody Wants to Run
The staffing crisis does not exist independent of who is inside the facilities. This is the question the Safe Prisons Initiative does not address, and the silence on it is the most consequential omission in the entire five-component framework.
Michigan’s prisons are operating at 95.5% of net capacity. At the same time, half of the facilities are running with vacancy rates between 15 and 35%. That combination — near-capacity population, dramatically below-capacity staffing — is not manageable through technology. It is a ratio problem, and ratios are solved by changing the numerator, the denominator, or both.
Nationally, at least 128,000 people are incarcerated for non-criminal technical violations of probation or parole, accounting for 27% of all state and federal prison admissions, according to the Prison Policy Initiative’s 2025 analysis. These are people who violated supervision conditions, which can include missing a meeting with a parole officer, failing a drug test, or violating curfew, not people convicted of new criminal offenses. They occupy the same beds as violent offenders, require the same administrative processing, consume the same medical and programming resources, and contribute to the population pressure that makes the officer-to-inmate ratio unsustainable.
Michigan has made real progress on this specific pipeline. Probation violators sent to prison declined 72% from the 2002 peak. Technical parole violators returning to prison are at a 40-year low, down 74% from the 2002 high. These are not small numbers. They represent a significant reduction in unnecessary incarceration, and the staffing crisis would be considerably worse without them. But they also represent the outer boundary of what has been politically achievable without addressing the sentencing structure, the absence of good time credits, and the lack of community-based supervision alternatives that would allow the system to right-size the population against the staff it can actually sustain.
Byron Osborn of the Michigan Corrections Organization did not call the Safe Prisons Initiative’s first problem to be scanner deployment or mail policy. He called it prisoner classification, specifically the practice of housing high-risk individuals in lower-security settings that were not designed for that population. This is a population management and risk-stratification problem, not a technology problem. Housing mixed-risk populations in facilities calibrated for lower-level supervision creates the exact conditions under which incident rates climb: power imbalances, increased vulnerability for lower-risk individuals, complex unit dynamics, and staff who are managing populations across a wider risk spectrum than their environment and staffing ratio were designed to handle.
The Fiscal Architecture of Doing It Wrong
The cost logic of the current approach is worth making explicit, because it does not show up as a single budget line anywhere in MDOC’s public reporting.
Michigan spends $49,290 per incarcerated person per year at the state level. For facilities with 18% or higher vacancy rates, that base cost is augmented by overtime expenditures. The Michigan Corrections Organization documented more than 544 mandatory overtime shifts at a single facility over 18 days in summer 2024. Overtime pay, workers’ compensation claims from injury in understaffed and high-tension environments, recruitment and onboarding costs for officers who leave within the first year, and litigation exposure from unsafe conditions are all downstream costs of the staffing crisis. These are not incidental expenses. They are structural outputs of a system that chose to address declining staff retention through technology investment rather than compensation and working condition reform.
A genuine safety initiative would ask: which people in these facilities do not require high-security containment and could be managed through community-based supervision at a fraction of the cost? Which officer compensation and scheduling structures would reduce voluntary departures below the current rate that equals or exceeds annual hiring? What prisoner classification changes would reduce mixed-risk housing and the incident rates that flow from it? And what does a sustainable officer-to-inmate ratio actually look like at each security level, and how does current staffing compare? Those are the questions a systems operator would ask. The Safe Prisons Initiative answered different questions, and the officers who work inside the facilities know the difference.
What Gets Ignored When the Narrative Is Technology
There is a pattern in how institutional systems respond to crises that have politically difficult root causes. They identify a component of the problem that is visible, quantifiable, and actionable without structural change, and they announce progress on that component. Contraband is visible. It can be photographed, counted, and reported. Scanners produce metrics. Mail policy changes generate press releases.
Officer burnout is not visible in the same way. A resignation is a row in a spreadsheet. The fact that the person who resigned had eight years of unit-specific knowledge that is now gone, and that the mandatory overtime burden on the remaining officers just increased by a calculable amount, and that the next incident in that unit will be handled by someone with fewer relational tools and less situational awareness, does not appear in a press release about contraband interception rates.
The research literature on correctional safety has been consistent on this for decades. A 1983 paper warned that officer stress and burnout lead to soaring costs from absenteeism and turnover. A 2000 analysis called a 16% annual correctional officer turnover rate “alarmingly high.” In Michigan in fiscal year 2023, 723 of 1,167 employee departures were correctional officers. The ratio of officer departures as a share of total separations rose from 52% to 61% in two years. These are not new warning signs. They are the same warning signs that have been compounding for decades, responding predictably to the same inadequate fixes.
The Safe Prisons Initiative is not wrong. Contraband control matters. Programming matters. Classification matters. But it is a partial response to a systems failure that requires a systems answer. Michigan is trying to stabilize a human infrastructure crisis with technology, while the people who constitute that infrastructure are describing, in direct and documented terms, exactly what would actually help. Until the state addresses officer compensation, working conditions, and population composition with the same urgency it applies to mail scanning policy, the assault numbers will continue to track upward, and every new initiative will arrive after the last one failed to move the needle on the variable that actually matters.
Sources and Documentation
Rita Williams, Scanner vs. Staff: Why the MDOC’s Safe Prisons Initiative Can’t Fix a Systems Personnel Failure, Clutch Justice (Apr. 16, 2026), https://clutchjustice.com/mdoc-safe-prisons-staffing-audit/.
Williams, R. (2026, April 16). Scanner vs. staff: Why the MDOC’s safe prisons initiative can’t fix a systems personnel failure. Clutch Justice. https://clutchjustice.com/mdoc-safe-prisons-staffing-audit/
Williams, Rita. “Scanner vs. Staff: Why the MDOC’s Safe Prisons Initiative Can’t Fix a Systems Personnel Failure.” Clutch Justice, 16 Apr. 2026, clutchjustice.com/mdoc-safe-prisons-staffing-audit/.
Williams, Rita. “Scanner vs. Staff: Why the MDOC’s Safe Prisons Initiative Can’t Fix a Systems Personnel Failure.” Clutch Justice, April 16, 2026. https://clutchjustice.com/mdoc-safe-prisons-staffing-audit/.
