Brief written August 23, 2026 from GPS Intelligence System data.(K)

Solutions: A Path Forward for Georgia's Prisons

Every other page in this intelligence system documents what is broken: at least 1,861 deaths in GDC custody since 2020 (GPS-tracked mortality data — a floor, since GDC reports deaths roughly two months late), a prison homicide rate nearly triple the national average, a correctional-officer workforce at 52.5% vacancy, and a federal finding that Georgia is "deliberately indifferent" to the violence in its own prisons (DOJ findings report, Oct. 2024).

This page is the counterpart. It documents what would fix the system — every proposal drawn from GPS's own reporting and research library, each with a cost, an evidence base, and a named obstacle. The through-line is simple: the crisis inside Georgia's prisons is not a mystery in search of an answer. It is a set of policy choices, and policy choices can be reversed. Several of the most powerful fixes cost little or nothing. One of them has been Georgia law since 1897.

1. The Flagship Fix: Separate the Gangs — It Costs Nothing, and It's Already the Law

The single cheapest, highest-leverage intervention available to Georgia is a classification and housing decision: house rival gangs apart from each other, and house gang members apart from the unaffiliated majority. No construction. No new appropriation. No new statute. Just enforcement of a duty Georgia wrote into its own code more than 125 years ago.

The problem this solves

Georgia has identified 315 different gangs operating inside its prisons and validated roughly 15,200 people — 31% of the entire incarcerated population — as gang-affiliated, more than double the national average of about 13% (GDC Security Threat Group data, cited in GPS, "315 Gangs, Zero Strategy"). The Commissioner himself told the 2024 Senate Study Committee that about 31% of the population are validated STG offenders (2024 Georgia Senate Study Committee final report). Yet Georgia has no systematic gang-separation housing policy, no structured gang-renouncement program, and no operational strategy for keeping rival factions apart (GPS, "315 Gangs, Zero Strategy").

The body count from that choice is documented. DOJ tracked GDC homicides climbing from 7 in 2018 to 35 in 2023 (DOJ findings report); independent counts documented at least 66 homicides in 2024 (GPS research library, Staffing Crisis & Correctional Officer Turnover). GPS has tracked at least 260 homicides in custody since 2020 (GPS-tracked mortality data). In the first months of 2026 alone, GPS documented: four men killed in a single gang disturbance at Washington State Prison on January 11 — including Jimmy Trammell, who had 72 hours left on a ten-year sentence, in a facility running five officers against 69 posts (GPS, "Separate the Gangs or Keep Burying the Dead"); a coordinated system-wide gang eruption on April 1 that locked down every state prison, dispatched life flights to two facilities, and confirmed stabbings at five (GPS, "Blood on Blood"); and another statewide lockdown on May 21 after deaths at Ware and Augusta (GPS, "Separate the Gangs. It Costs Nothing.").

A statewide lockdown, as GPS's reporting puts it, is not a strategy. It is a confession — an admission that the state cannot say who is safe in which dorm, because it has spread the same rival gangs across every fence in Georgia.

The law Georgia refuses to enforce

Here is what makes this fix uniquely actionable: it requires no new law. Under O.C.G.A. § 42-5-52, tracing to 1897, the department "shall provide for the classification and separation of inmates" — including separating habitual criminals and "incorrigibles" from the rest (GPS, "On the Books Since 1897: The Separation Law Georgia Refuses to Enforce"). A second Georgia statute declares every person's right to be safe from the fear and violence of gangs. The department honors neither. GPS has separately documented four medium-security prisons quietly packed with close-security men — a classification failure running in exactly the wrong direction.

The state is not missing a tool. It is ignoring one it has held since the nineteenth century.

The evidence that separation works

Peer states faced the same crisis decades ago and built documented responses (GPS, "315 Gangs, Zero Strategy"):

• Texas developed the GRAD (Gang Renouncement and Disassociation) program, with 2,668 completions as of 2012 (GPS Comparative Solutions Evidence Base). Caveat: no rigorous recidivism or before/after violence evaluation of GRAD exists — the leading academic study explicitly declines to answer whether renunciation works. The architecture is proven as policy; the outcome data is a documented gap GPS has flagged for open-records follow-up.
• Arizona ran an STG step-down program; its own data showed gang members committed disciplinary violations at two to three times the rate of other prisoners — the core empirical case for separation. Caveat: raw before/after incident counts were not located; GPS has flagged ADCRR data for request.
• California moved more than 1,512 people out of solitary under the Ashker settlement and cut Pelican Bay's long-term isolation population from 513 to 2 — a 99.6% reduction with no reported violence surge (GPS Comparative Solutions Evidence Base).
• Northern Ireland separated paramilitary factions outright. Caveat: outcome data not located in GPS's research; the model exists as precedent, not as a measured result.

Cost, feasibility, and the obstacle

Cost: effectively $0. Separation is a housing-assignment and classification decision. It requires no new beds and no new money — GPS's title says it plainly: "It Costs Nothing. Georgia Keeps Choosing the Bodies." If anything, separation saves: solitary confinement "requires twice as many guards," and every gang-driven lockdown burns overtime, medical flights, and emergency response (GPS Comparative Solutions Evidence Base).

The obstacle is GDC itself. This is not a funding problem or a legislative problem; it is a leadership choice. The department validates 15,200 gang members, knows which sets are at war, and manages the result with bulletins and lockdowns instead of housing policy. The pressure point is therefore public and political: the General Assembly can compel enforcement of § 42-5-52 through oversight hearings or a clarifying statute, and the Governor can direct it tomorrow.

2. Vision 2027: The Legislative Package

Vision 2027 is GPS's post-conviction justice reform framework — a legislative package synthesizing more than a dozen GPS research collections into specific statutory levers. Bill-level text lives on the framework page; the planks below are the ones GPS's research grounds most heavily, each with its documented lift and its documented obstacle.

Parole reform and lifer review

Georgia's parole board has functionally stopped saying yes to lifers: in FY2025 it granted release to 87 lifers serving for serious violent felonies out of 2,277 considered (BOPP Annual Report FY2025, cited in GPS, "The Board That Stopped Saying Yes"). Average time served on a life sentence rose from under nine years in 1973 to 28 years by FY2022 (BOPP FY2022). This is fiscally irrational on the state's own numbers: parole supervision costs $4.65 a day against $49.35 for incarceration (BOPP FY2010), parolees complete supervision at ~70% — well above the national average (BOPP FY2023) — and parole supervision avoided the state $319 million in FY2023 alone (BOPP FY2023). Obstacle: the parole board's risk aversion and opaque set-off practices — GPS reporting documents families and incarcerated people preparing §1983 complaints against the board alleging "arbitrary and capricious set-offs" (GPS claims layer, family attestations).

Geriatric and compassionate release

Georgia holds 5,753 people aged 60 or older (GPS, End the Warehouse). The research case is overwhelming: the U.S. Sentencing Commission found older offenders' rearrest rate (21.3%) is less than half that of under-50 offenders (53.4%); the federal CARES Act released 11,000+ elderly and medically vulnerable people with a 0.15% new-arrest rate; compassionate-release recidivism runs 3.5% versus 41% for the general federal population; California's elderly parole re-conviction rate is 1.8% (GPS research library, Aging Prison Population). GDC's own Aging-Inmate data puts medical cost at $8,500/year for prisoners 65+ versus $950 for younger prisoners — a 9x multiplier. Forty-five states plus the federal government have compassionate-release laws and barely use them. Obstacle (named in GPS research): the parole board's risk aversion and restrictive eligibility criteria.

Independent corrections oversight

Covered in detail in Section 5 — a statutorily independent ombuds office on the New Jersey/Washington model, roughly $2.8 million a year, against a settlement floor of $50,633,556 across 261 claims (GPS, "What Georgia Pays," compiled from Georgia DOAS risk-management records and court files).

Death-reporting transparency

Covered in Section 7 — a statutory requirement that in-custody deaths and causes be reported publicly and promptly, reversing GDC's March 2024 blackout.

Correctional education funding

Georgia spends roughly $3.44 per incarcerated person per year on vocational education ($172,000 against ~50,000 people) and prohibits incarcerated students from state financial aid (Brennan Center, March 2026, cited in GPS research). The RAND meta-analysis finds correctional education participants are 43% less likely to recidivate, with a return of $4–$5 per dollar invested (GPS Budget Baseline FY2025–FY2027). Obstacle: budget allocation, not evidence.

Presumptive field-test reform

Georgia is the only state in the nation where unconfirmed colorimetric field drug-test results remain admissible at trial for non-marijuana cases (Quattrone Center, cited in GPS research). Field-test error rates run as high as 38% in some contexts; New York City's independent lab retests found an 85% false-positive rate for fentanyl field tests; Colorado's own DOC measured a ~33% false-positive rate. Colorado's fix — HB 26-1020, requiring summons rather than arrest and confirmatory lab testing — passed unanimously in both chambers with a $0 fiscal note in March 2026 (GPS Comparative Solutions Evidence Base). Obstacle: none fiscal; purely legislative initiative.

Gang-separation enforcement

The § 42-5-52 plank from Section 1 — codifying what 1897 already commanded, with reporting requirements so non-compliance is visible.

The lift: not one of these planks costs more than a rounding error against a $1.9 billion agency (GDC FY2025 actual spending: $1,913,888,054 — Governor's Budget Report FY2027, via GPS budget analysis). Several return money. The full case is assembled on the Vision 2027 page.

3. End the Warehouse: Rehabilitation That Pays for Itself

The End the Warehouse campaign starts from a diagnosis GPS documented across a five-part series: strip a prison of work, school, purpose, and family, and the vacuum fills with exactly what Georgia is drowning in — drugs, gangs, despair (GPS, "The Only Family Left"; "Reopen the Doors — Normalization"). The fix has a name — normalization: life inside should resemble life outside as closely as security allows, because roughly 95% of incarcerated people are ultimately released, and the point is to return them intact.

What Georgia buys instead

GPS's budget analysis documented a state that spends $120 million on surveillance technology and $2.6 million on rehabilitation programming — roughly $52 per person per year on programming (GPS, GDC Budget analysis). A $150 million surveillance hub (the OWL command center) went live in mid-2026 while $805,000 was left for classrooms; GPS's canonical line-item review computed a 22:1 surveillance-to-rehabilitation ratio across the AFY2026 amended and FY2027 approved budgets, even as the State Prisons program was cut $187.5 million (16.78%) from FY2025 to FY2027. Alabama — itself under federal oversight for unconstitutional prisons — still outspends Georgia 19-to-1 per inmate on education (GPS Budget Baseline FY2025–FY2027). Meanwhile Georgia's recidivism rate has sat near 30% for a decade despite doubled corrections spending (Georgia Council on Criminal Justice Reform, via GPS research).

The measured returns elsewhere

• RAND meta-analysis: correctional education cuts recidivism 43% and returns $4–$5 per dollar in avoided reincarceration costs (confirmed, GPS research library).
• Michigan Vocational Village: 2019 graduates recidivated 6.5 percentage points below the state rate (moderate confidence).
• Vera Institute's Restoring Promise (South Carolina, randomized trial): 73% reduction in the odds of violent incidents and 83% reduction in the odds of restrictive-housing stays (moderate confidence).
• "Little Scandinavia" at SCI Chester, Pennsylvania: a ~$310,000 renovation producing a unit run at a 1:8 officer-to-resident ratio (versus 1:128 in the rest of the facility), with a single physical altercation since opening in 2022 and almost no violence in 2024; Pennsylvania announced expansion to three more facilities in March 2025. Caveats: results are early and mostly non-causal — researchers stress this — and Pennsylvania's system overall saw a ~21.6% violence increase in 2024, meaning the unit is an island, not yet a system (GPS Comparative Solutions Evidence Base).
• Norway: reconviction around 18–20% at two years and 25% at five, down from a pre-reform 60–70%; Bastøy Prison runs 16% recidivism. Caveat (must travel with these numbers): Nordic outcomes reflect far shorter sentences, much smaller populations, and stronger social-welfare baselines. They establish that measurable alternatives exist — not that they transplant one-to-one to Georgia (GPS methodology note).
• DOJ meta-analysis of reentry programs: an average ~6% recidivism reduction, larger when treatment begins in prison and continues after release. Caveat: NIJ cautions no reentry program yet meets the replicated-RCT "evidence-based" standard; results are mixed.

Cost and obstacle: the transformation is predominantly a reallocation, not an appropriation — Georgia already spends at an all-time peak; it spends it on watching instead of teaching. The obstacle is the budget's center of gravity: surveillance and contract beds (~$140 million/year to CoreCivic and GEO Group for ~7,800 prisoners, at per-day costs roughly 10% above state prisons per a Georgia Department of Audits FY2018 analysis) crowd out the programming line every cycle. Changing that requires legislators to amend the GDC budget — which is why End the Warehouse is built as a public sign-on campaign, not a white paper.

4. Staffing and Medical Care: Pennsylvania Proved the Turnaround

Every documented failure in Georgia's prisons runs through the staffing collapse: 52.5% systemwide correctional-officer vacancy (2,985 of 5,991 budgeted positions vacant as of January 2024), 80% vacancy at Valdosta State Prison by April 2024 (DOJ), 82.7% of new hires gone within a year (DOJ), and dead-last, 50th-of-50 officer pay (GPS research library). Former GDC assistant warden Russell Zirkle testified on record that blended state averages hide 70–80% vacancy rates at major prisons — naming Telfair, Hancock, Smith, Valdosta, Macon, Autry, and Calhoun — and former officer Tyler Ryals corroborated the staffing cover-up (GPS corroboration backbone). GDC Commissioner Timothy Ward testified in 2022 that annual officer turnover was 49%. See the Staffing Crisis brief.

The solution is not hypothetical — it is measured. Pennsylvania created a dedicated Recruitment and Retention Division in late 2022, held 750+ job fairs and events in 2024, set a trainee starting salary of $46,986 rising to $49,156, and cut its officer vacancy rate from 10.5% to 4.8% in two years (GPS Comparative Solutions Evidence Base, moderate confidence). Alabama's March 2023 pay package (trainee salary near $57,000) made officers 28% less likely to resign — caveat: hiring collapsed anyway and total staff still fell 55% over nine years, proving pay alone, without a recruitment pipeline, is insufficient. North Carolina's governor proposed a 15% biennium raise; outcome data is pending.

Cost and obstacle: a dedicated recruitment division is an organizational expense in the low millions against a $1.9 billion budget; pay competitiveness is the larger but still marginal line — Georgia's settlement floor of $50.6 million would fund a substantial raise on its own. The obstacle is that Georgia has spent its increases elsewhere (see the surveillance allocation above) and remains #50 in pay by choice.

Medical care runs on the same staffing logic plus contractor accountability. The documented failures: GDC has no hepatitis C test result on file for 27,471 people — 51% of its custody population (GDC Inmate Statistical Profile, June 2026, via GPS research); the aging population grew from ~5% to ~27% in thirty years; and Legionella at Autry State Prison ran a four-year detection-to-disclosure lag (first GPS-confirmed case July 2018; first public acknowledgment October 2022), with remediation-by-flushing failing for 17+ months — consistent with national evidence that flushing and chlorination alone consistently fail (GPS Legionella research; CHCF Stockton spent $8.5M and still had 21 of 29 buildings restricted). The fix: mandatory communicable-disease testing and reporting, engineered water remediation with verification monitoring, and medical-contract performance auditing. The Medical Neglect brief assembles the evidence.

5. Oversight: The $2.8 Million Watchdog Georgia Refuses to Buy

Georgia is one of the states with no independent prison oversight body; roughly 20 states plus D.C. now have one, and the Brennan Center's March 2026 report names Georgia among the laggards (GPS research, Oversight & Investigations).

The models are costed and running:

• New Jersey's Office of the Corrections Ombudsperson — strengthened by the 2020 Dignity Act — is fully independent, reports to the Governor, holds subpoena power and unannounced-inspection authority, covers 9 prisons (~13,000 people), and costs $2,806,000 in FY2025 with 26 staff (GPS Comparative Solutions Evidence Base, moderate confidence).
• Washington's Office of Corrections Ombuds (created 2018, RCW 43.06C) oversees 11 prisons (~13,075 people); its statutory purpose explicitly includes reducing DOC litigation exposure, and its correspondence is confidential and privileged (confirmed).

The cost framing is decisive: New Jersey's entire office costs a fraction of a single conditions settlement — the Ashker plaintiffs' attorney fees alone exceeded $4.5 million. Georgia's own settlement floor stands at $50,633,556 across 261 claims (GPS, Legal Settlements). Design caveat from the research: the UK's HM Inspectorate of Prisons "cannot enforce" its findings — Georgia should pair inspection with an enforcement or mandatory public-response mandate, not copy a toothless model.

Obstacle (named in GPS research): neither GDC nor the Governor's office has supported an independent office; enabling legislation is required. That makes this a pure legislative ask — and one any member can carry.

6. Solitary: Fewer Isolation Cells, More Safety, Real Savings

Georgia's Tier programs and Special Management Unit have drawn federal findings, a contempt record, and litigation-documented conditions: cells "smaller than the average parking space," 22–24 hours a day of confinement, 78% of SMU prisoners held in isolation more than two years as of 2017, an average stay of three to four years, and 70 of 180 SMU prisoners designated mentally ill — housing expert Dr. Craig Haney called "dangerous" (Gumm v. Ford record, via GPS, Solitary Confinement). New York City jail data shows why this kills: people with any solitary exposure were 7.3% of admissions but accounted for 53.3% of self-harm acts (Kaba et al.).

The reductions are among the best-measured results in American corrections:

• North Dakota: a 74.28% reduction in solitary use (2016–2020) — confirmed, peer-reviewed — with monthly solitary sanctions falling 99% at JRCC and 59.1% at the state penitentiary.
• Oregon's Resource Team: among participants (averaging 9.7 prior solitary admissions), disciplinary infractions fell 55.7% and assaults 73.9%; Oregon's Behavioral Health Unit cut staff use-of-force nearly 86% (2016–2021).
• California post-Ashker: statewide SHU population down 65% (9,870 → 3,471), Pelican Bay long-term isolation down 99.6%, no reported violence surge.

Cost: negative. Solitary "requires twice as many guards" (GPS Comparative Solutions Evidence Base) — every isolation bed closed frees officers in a system missing half of them. Obstacle: GDC uses the Tier programs as its default control strategy in the absence of classification and staffing; change has historically arrived only through litigation (Gumm), which is the most expensive possible path — for the state and for the people inside.

7. Transparency: Count the Dead, Feed the Living, Fix the Water

Death reporting

In March 2024, GDC stopped providing cause-of-death information in its monthly mortality reports, deferring to local processes that can run indefinitely (first reported by the Atlanta Journal-Constitution; GPS research, Deaths in Custody). Of the at least 1,861 deaths GPS has tracked since 2020, 688 carry no known cause at all (GPS-tracked mortality data). The structural enablers are documented: in 155 of 159 counties, death certification belongs to an elected coroner with no medical training requirement (O.C.G.A. § 45-16-1), and the GBI Medical Examiner's Office runs about 20% understaffed (15 of 19 budgeted examiners). When GPS obtained Ware State Prison death-investigation records through an open-records request, they documented at least 19 homicides at a single prison — 11 naming a suspected attacker — with no public record of prosecution in any of them (GPS, "At Least Nineteen"; Ware State Prison).

The fix: a statute requiring prompt public reporting of every in-custody death and its manner and cause, full funding of the GBI Medical Examiner's office, and autopsy standards that don't depend on a family's ability to hire a private pathologist. Cost: low millions. Obstacle: GDC's demonstrated preference for opacity — the blackout was a policy decision, and reversing it requires legislation.

Food and sanitation

Georgia spends $0.54 per meal — 14.8% of the American Correctional Association's ~$3.66 benchmark; meeting the benchmark would cost roughly $211 million/year, 6.8x current spending (GPS, Prison Nutrition). The inspection regime masks the reality: GPS's review of DPH food-safety data found most facilities scoring 90–100 while the three failing inspections documented rats and roaches in kitchens (Johnson SP, 64), a smashed handwashing sink and sewage backing through floor drains (Pulaski SP, 67 — back to 96 eight days later), and rodent activity in every Smith SP inspection from 2022 through 2025 (Scores Without Sanitation). The score measures preparation for inspection day, not conditions. The fix: fund food operations to benchmark, unannounced inspections, and published corrective-action verification. Obstacle: appropriations priority — the money exists in a $1.9 billion budget; it is spent elsewhere.

8. The Master Finding: Fewer Prisoners, Less Crime — Everywhere It's Been Tried

The best-documented finding in GPS's Comparative Solutions Evidence Base is decarceration without crime increases:

• New York more than halved its prison population (1999–2023) while violent crime fell 34% — faster than the national 28% decline — and closed more than a dozen prisons. New York City (1996–2014): serious crime down 58%, incarceration down 55%, 31,120 fewer residents behind bars.
• New Jersey holds 37% fewer people than in 2019. Alaska, Connecticut, New Jersey, New York, and Vermont have each cut prison populations over 50% from peak.
• California downsized 23% (2006–2012); violent crime fell 21%.
• The Netherlands cut its prison rate ~46% (2005–2016) and closed roughly half its prisons as crime fell. Caveat: European transferability limits apply, as with the Nordic figures.
• Nationally, the U.S. prison population fell 25% (2009–2021) while reported violent crime dropped to half its 1990s level.

Methodology caveat (applies to every recidivism figure cited on this page): definitions and windows differ across all sources — rearrest vs. reconviction vs. reincarceration; 2-, 3-, 5-, and 8-year windows — so raw cross-jurisdiction comparisons should be read as directional, not interchangeable.

Georgia's levers to capture this are the documented ones above: parole that functions (the state's own board quantifies the savings — $319M in FY2023), geriatric release of the statistically safest cohort in custody (21.3% rearrest; 0.15% CARES new-arrest rate), and sentencing-integrity fixes like field-test reform. And the fine/fee evidence shows debt discharge often costs governments nothing in realized revenue — eight reforms across six states lifted $2.6 billion in unpayable debt.

9. What You Can Do — For Each Fix

Separate the gangs ($0, no legislation required): Share GPS's case — "Separate the Gangs. It Costs Nothing. Georgia Keeps Choosing the Bodies." and "On the Books Since 1897" — with your state representative and senator, and ask them directly: will you demand enforcement of O.C.G.A. § 42-5-52? Find them at gps.press/find-your-legislator.

Vision 2027 (legislative): Sign on to the campaign and read the full framework at gps.press/intelligence/issue/vision-2027/. Ask your legislators which plank they will carry — oversight ($2.8M), death-reporting transparency, geriatric release, education funding, or field-test reform ($0 fiscal note, unanimous in Colorado).

End the Warehouse (budget): Add your name at gps.press/intelligence/issue/end-the-warehouse/ and press the budget question: why $150 million for a surveillance hub and $805,000 for classrooms, when RAND measures $4–$5 back on every education dollar?

If you have direct knowledge: families and incarcerated people can report conditions at gps.press/submit-a-report/ or share a story through Tell My Story. The mortality archive at gps.press/georgia-prison-deaths/ exists because families spoke.

Reporters: the story leads and documents are organized in the media brief. Advocates: the action map is the advocate brief. Legislators and staff: the fiscal exposure analysis is the legislative brief.

The Bottom Line

Georgia does not lack options. It lacks only the decision to use them. The cheapest fix — separating the gangs — has been statutory law since 1897 and costs nothing. The oversight office costs $2.8 million against a $50.6-million settlement floor. The education investment returns four to five dollars per dollar spent. The parole and geriatric-release levers free the statistically safest people in custody while the state's own board books hundreds of millions in avoided costs. Every one of these has a named model, a measured outcome, and a named obstacle.

Do three things now: (1) sign on to Vision 2027 and End the Warehouse; (2) contact your legislator and ask one question — will you co-sponsor independent oversight and demand enforcement of the 1897 separation law?; (3) share the gang-separation case with one person who doesn't know it. The state has been choosing the bodies. The record above shows it could choose otherwise — starting this session.