Brief written August 2, 2026 from GPS Intelligence System data.(K)

Legislative Intelligence Brief: Georgia Department of Corrections — Fiscal Exposure, Legal Liability, and the Reform Window

Prepared for: Members of the Georgia General Assembly, legislative staff, and committee analysts Prepared by: Georgia Prisoners' Speak (GPS) Intelligence Desk

Executive Summary

Georgia now operates the most expensive prison system in its history — $1.79 billion in the approved FY2027 budget (HB 974) — atop a correctional apparatus that the U.S. Department of Justice has formally found to violate the Eighth Amendment. GPS has independently tracked at least 1,859 deaths in GDC custody since 2020, a figure that is a floor, not a total, because the state reports causes of death roughly two months late and leaves more than half of all deaths classified as unknown or pending.

The liability picture is quantifiable. GPS's compilation of Georgia Department of Administrative Services risk-management records and court files places prison death and injury settlements at a floor of $50,633,556 across 261 claims since 2008, with the DOJ's October 2024 findings letter and its 82 recommended remedial measures projecting substantially larger exposure if a CRIPA enforcement action or consent decree follows. Meanwhile, the budget itself documents the state's priorities: roughly $27.1 million in new security and surveillance appropriations against $1.23 million for rehabilitation across the amended FY2026 and FY2027 cycles — a 22-to-1 ratio — and a food budget of $0.54 per meal, under 15 percent of the American Correctional Association benchmark.

This brief organizes the evidence around what the General Assembly can act on: settlement exposure, pending litigation, mortality and tracking gaps, staffing data integrity, sanitation inspection data, the gang-separation duty Georgia has carried in statute since 1897, the Vision 2027 post-conviction package, rehabilitation return-on-investment, and specific bill-language entry points. A district-relevant facility directory closes the brief. Every figure traces to GDC's own budget documents, court records, federal findings, county health inspections, or GPS's open-records productions.

1. The Fiscal Baseline: A $1.8 Billion Agency Reallocating Away From Rehabilitation

Topline and trajectory

GDC consumed $1,913,888,054 in actual FY2025 spending — an all-time peak — after a 57 percent surge in just four years off a plateau of roughly $1.2 billion that had held for most of a decade (GPS budget analysis, drawn from the Governor's Budget Report FY2027). The amended FY2026 budget stands at approximately $1.80 billion, and the FY2027 approved budget under HB 974 at roughly $1.79 billion. The single-year FY2024-to-FY2025 jump alone was $387 million, the largest in the visible window.

As of GDC's July 31 population snapshot, 52,894 people were in custody: 34,965 in state prisons, 8,278 in private prisons, 4,052 in county prisons, 2,707 in transitional centers, and the remainder in probation RSAT and detention facilities — with 2,542 state inmates backlogged in county jails awaiting transfer. That backlog is a direct county-budget pressure in members' districts: local jails house state inmates on state responsibility.

The composition of spending matters more than the topline. Between FY2025 actual and FY2027 approved, GDC's own budget documents show the State Prisons program cut by $187 million (–16.8%), while Private Prisons grew $20.9 million (+13.7%), the Health program added $37.3 million (+9.6%), and Offender Management climbed $8.1 million (+16%). GDC administrative headquarters took an 18.4 percent cut. The state is paying more than ever and operating less of the system itself.

The 22-to-1 surveillance-to-rehabilitation ratio

Across the amended FY2026 and FY2027 cycles, GPS's line-item analysis identifies approximately $27.1 million in new security and technology appropriations against $1.23 million for rehabilitation — a 22-to-1 ratio. The security investments include at least $35 million for managed-access cell-phone blocking, $7.2 million for body cameras and Tasers, $6.9 million for the OWL surveillance unit plus another $6.9 million in FY2027 OWL technology and personnel, $4.1 million for digital forensics, $2.5 million for officer tablets, roughly $1 million for off-site mail screening, and $84.7 million in multi-year capital commitments for thermal cameras, CCTV, and perimeter security.

The rehabilitation side of the ledger, in full: $93,672 for high school diploma program staff in amended FY2026; $849,033 in FY2027 — $104,000 below the Governor's own request, after the legislature directed GDC to "explore virtual high school options"; $243,179 in reentry programming in amended FY2026, including a $150,000 peer-led pilot at Autry State Prison that was not continued into FY2027; and $93,179 in one-time Metro Reentry Facility programming. Total new rehabilitation investment: approximately $26 per incarcerated person per year.

The RAND Corporation's meta-analysis found correctional education reduces recidivism by 43 percent and returns four to five dollars per dollar invested; GDC's own vocational completers recidivate at about half the general population rate. The state is declining to spend on what its own data says works.

Food, health, and hidden cost shifts

GDC's FY2027 Food and Farm Operations line is $31.3 million — $1.62 per person per day, or $0.54 per meal (GPS prison nutrition analysis, from GDC budget documents). The ACA benchmark is approximately $3.66 per meal; meeting it would cost roughly $211 million a year — nearly seven times current spending. GDC's own SOP 409.04.02 mandates only two meals on weekend days. In 2015, the Aramark contract at Hays and Smith State Prisons provided $2.97 per prisoner per day — meaning the real per-meal figure has fallen by nearly half over the decade. Meanwhile, health costs have surged past $432 million, a cost trajectory consistent with chronic underfeeding, deferred care, and deteriorating conditions shifting expense into the medical line.

One more revenue stream deserves appropriations-committee attention: GDC's 2016 Securus telephone contract paid the department a $4 million one-time incentive plus a 59.6 percent commission on call revenue with a $325,000 monthly floor — roughly $8 million a year flowing back to the agency while families paid approximately 13 cents per minute. These commissions pool in opaque "Inmate Welfare Funds." (GPS End the Warehouse analysis.)

2. Settlement and Verdict Exposure

The ledger: GPS's compilation of Georgia DOAS Risk Management settlement records and court files — spanning 2008 through 2025 — places GDC death- and injury-linked payouts at a floor of $50,633,556 across 261 claims (legal settlements analysis). Individual payouts range from a few hundred dollars to multi-million-dollar settlements; payouts linked to deaths form a particularly damning cluster, and the 2023 payment year was among the heaviest in the ledger.

Three features of this record matter for fiscal analysis:

1. The floor undercounts harm. The $50.6 million reflects only claims that survived to payment. GPS's litigation tracker shows recently terminated federal civil actions captioned against GDC officials — e.g., Grant v. Ward (GAMD, 5:22-cv-00396), Humphreys v. Oliver (GAND, 1:25-cv-07012), Daker v. Oliver (GAND, 1:25-cv-03191) — all closing with $0 recovery. This is not evidence that the underlying harms are small; it is evidence that the Prison Litigation Reform Act's exhaustion requirement and GDC's own grievance architecture (Section 10, below) filter most claims out before any fact-finding. Liability concentrates in the deaths and catastrophic injuries that cannot be grieved away.

2. Court-verified records show institutional resistance to accountability. A federal court has held GDC in contempt for falsified records, with one judge stating the system had "no desire or intention to comply" with court orders (violence and mental health analyses). Contempt findings multiply litigation costs: they extend monitorships, invite fee awards, and convert defensible cases into indefensible ones.

3. The claims pipeline is being suppressed, not resolved. GPS has received family-attestation accounts — for example, from Ware State Prison — alleging the state provided no answers about a wrongful death until the statute of limitations expired in 2024. Whether or not each such account is verifiable, the pattern they describe is a fiscal time bomb: suppressed claims do not disappear, they arrive later, with worse facts and higher punitive exposure.

3. Pending Litigation and Projected Liability

The dominant risk is the federal record. The U.S. Department of Justice's October 2024 findings report — 93 pages, produced after what DOJ described as an "unnecessarily contentious" records fight in which GDC refused to produce documents until a subpoena was enforced and conducted pre-inspection cleanups — found reasonable cause to believe Georgia violates the Eighth Amendment by failing to protect medium- and close-security prisoners from violence and LGBTI prisoners from sexual abuse. DOJ documented 142 homicides from 2018–2023 (rising from 7 in 2018 to 35 in 2023), more than 1,400 violent incidents in a sixteen-month period across 24 prisons, and staffing vacancy rates peaking at 60 percent systemwide. The report recommends 82 remedial measures and notes the Attorney General may sue under the Civil Rights of Institutionalized Persons Act. As of GPS's May 2026 research, no CRIPA enforcement action had been filed, and the posture of the current federal administration is uncertain (oversight analysis).

For legislative planning purposes, the CRIPA question is almost beside the point. The findings letter is a roadmap any future administration — or any private plaintiff's counsel citing it — can execute. States that have gone through CRIPA consent decrees know the cost profile: years of monitorship, court-supervised staffing and medical benchmarks, and attorney-fee exposure dwarfing the remedial spending the legislature declined to appropriate up front. Every remedial measure Georgia funds now is a remedial measure it will not fund later at consent-decree prices with a monitor's overhead attached.

Additional pending exposure categories:

• Restrictive housing. Court-verified litigation over the Special Management Unit — brought by lead plaintiff Timothy Gumm, held continuously for more than seven years — has already forced reforms; as of July 2017, 44 percent of the SMU population had been confined more than four years and 26 percent more than five, with Robert Watkins and Johnny Mack Brown each held eight to ten years (solitary confinement analysis). Prolonged-isolation claims remain a live liability category under evolving Eighth Amendment doctrine.
• Sexual abuse. DOJ found sexual assault "rampant"; GDC's own consultants, PREA Auditors of America, reviewed 388 PREA investigation files in May 2022 and found not one met statutory standards — yet all 273 facility-level PREA audits across five cycles returned 100 percent compliance. Of 15,542 PREA allegations from 2014–2024, only 3.5 percent were substantiated (sexual abuse analysis). This gap between paper compliance and investigated reality is precisely what plaintiffs' experts are paid to exploit.
• Wrongful death at scale. GPS's open-records production of Ware County coroner and GBI records — 231 death-investigation reports covering 84 Ware State Prison deaths since 2020 — documents at least 19 homicides at a single prison, with a suspected attacker named in eleven and no public record of prosecution in any. Each of those deaths is a potential wrongful-death action; the records establishing manner of death already exist in the state's own files ("At Least Nineteen: The Murders the State Didn't Prosecute," GPS, July 2026).

4. Mortality Trend and the State's Tracking Gap

GPS has independently tracked at least 1,859 deaths in GDC custody since 2020 (GPS-tracked mortality data, n=1,859+ — a floor, because GDC reports deaths on a delay and does not publicly release timely cause-of-death information). The annual progression:

• 2020: at least 294 | 2021: at least 257 | 2022: at least 254 | 2023: at least 262 | 2024: at least 333 — the deadliest year on record, 27 percent above 2023 | 2025: at least 301 | 2026 year-to-date: at least 158

Of the total, 1,044 deaths — more than half — remain classified as unknown or pending cause; only 417 have been attributed to natural or medical causes. GPS's tracking includes at least 256 confirmed homicides since 2020. Where the state attributed 66 deaths to homicide in 2024, GPS documented at least 100. In the first twenty days of July 2026 alone, GPS recorded at least eight deaths across six facilities: Central, Pulaski, Valdosta, Baldwin, Augusta State Medical Prison, and Burruss CTC. (Deaths in custody analysis; GPS mortality archive.)

The tracking gap is itself a policy failure the legislature can fix. GDC publishes no official prison-homicide count. DOJ found GDC's incident reports "routinely misclassified," with homicides listed as unknown and deaths following months of untreated illness recorded as "natural." GPS's July 2026 report "The Empty Column: Georgia Won't Say How Prisoners Die" documents the data void directly. In one Telfair State Prison case, GPS's analysis of records obtained through open-records requests found discrepancies between the autopsy report and GDC's own records, including conflicting information about location and the reported involvement of another person; separately, GPS analysis indicates a medical examiner's manner-of-death determination relied on GDC reports stating no foul play. GPS has also received family-attestation accounts disputing suicide classifications — including one at Ware — that families believe conceal other causes. These are not fringe claims; they are the predictable output of a system in which the agency under investigation supplies the foundational facts for the death investigation.

Geriatric mortality is a budget issue. GDC's own statistical profile counts 5,753 people aged 60 or older in custody. Recent GPS-tracked deaths include men aged 85 (Ware, June 2026), 77 (Coastal, June 2026), 66 (GDCP, June 2026), 65 (Wilcox, June 2026), and 62 (ASMP, June 2026). Georgia pays ICU prices to warehouse people old enough to draw Social Security, in facilities the DOJ says cannot keep anyone safe.

5. Staffing: Concealed Vacancy Numbers With Direct Fiscal Consequences

GDC's own consultants, Guidehouse (retained by the Governor), described a system in "emergency mode": 52.5 percent systemwide correctional officer vacancy as of 2024, 20 of 34 state prisons at emergency levels, eight facilities above 70 percent, and Valdosta State Prison at 80 percent — the facility housing the highest concentration of gang members and people with serious mental illness. The Guidehouse report, obtained by GPS through an open-records request, documented night shifts where one or two officers are responsible for an entire prison (staffing crisis analysis).

The official numbers understate it — on the record. Russell Zirkle, a former GDC assistant warden with a career spanning Valdosta and Autry, testified on-record and timestamped that the state's vacancy figures are "skewed" by blending fully staffed halfway houses into system averages, concealing 70–80 percent vacancy rates at major prisons — naming Telfair, Hancock, Smith, Valdosta, Macon, Autry, and Calhoun. Tyler Ryals, a former GDC officer, gave corroborating testimony; Zirkle's response: "Everything he says is completely true." And a GPS open-records production confirmed the operational reality: six officers on duty at Washington State Prison when the January 11, 2026 riot began — a facility running roughly 72 percent vacancy, with five officers covering 69 posts, where four men were killed that day, including Jimmy Trammell, who had 72 hours left on a ten-year sentence.

The pay response has not worked. Starting pay of $40,000–$43,000 ranks Georgia 50th of 50 states for correctional officer compensation (ZipRecruiter), and GPS's analysis of the promotion pipeline ("The Game They Learned," May 2026) traces how a closed, promote-from-within command culture produced the May 13, 2026 Tattnall County indictment of former Smith State Prison Warden Brian Adams on racketeering, bribery, evidence-tampering, and oath-violation charges — prosecutors allege he took bribes from a prison gang and buried shanks and contraband phones in his backyard. The legislature has doubled pay; the vacancy rate has not moved. The next dollar belongs in structure — separation, staffing-definition honesty, and oversight — not in repeating the failed raise.

6. Conditions and Sanitation: DPH Data the State Doesn't Aggregate

Georgia already inspects every prison kitchen. County sanitarians score them on a 100-point scale under O.C.G.A. § 26-2-370 and DPH Rule 511-6-1, and GDC's own SOP 409.04.26 confirms kitchens operate under local health department permits. No agency centrally collects or analyzes the results. GPS did (Scores Without Sanitation):

• Most facilities earn A grades; Central State Prison scored perfect 100s twice in 2025. Adequate food safety is demonstrably achievable inside GDC.
• Three prisons have failed since 2022: Johnson State Prison (64, December 2023) — rats and roaches throughout, rodent-gnawed bulk food with droppings and urine contamination, five broken ovens, holes in floors and ceilings, cold-holding violations; Pulaski State Prison (67, January 2026) — the culmination of a steady decline (83 → 73 → 78 → 67), with the single handwashing sink ripped from the wall, sewage backing up through floor drains as a repeat violation, and hot-holding failures (nacho meat at 65°F); Smith State Prison (68, May 2022).
• Pulaski's follow-up inspection seven days later scored 96 — a 29-point swing, which is precisely the problem: the scores measure a scheduled walkthrough, not what lands on the tray. GPS's reporting documents the gap between inspection-day conditions and daily practice, corroborated by incarcerated kitchen workers' accounts.

The legislative fix requires no new apparatus: DPH already inspects. A one-paragraph reporting mandate would make the existing data visible. Conditions more broadly — and their constitutional framing — are catalogued in GPS's facility conditions briefing.

7. Gang Violence and the Separation Statute Already on the Books

Georgia has identified 315 gangs operating inside its prisons and validated roughly 15,200 people — 31 percent of the incarcerated population — as gang-affiliated, more than double the national average of about 13 percent. It has no systematic gang-separation housing policy, no structured renouncement or exit program, and no dedicated operational strategy for keeping rival factions apart ("315 Gangs, Zero Strategy," GPS, March 2026).

The cost of that policy vacuum is documented in the state's own files. On January 11, 2026, four men were killed at Washington State Prison. On April 1, 2026, coordinated gang violence locked down every state prison, dispatched life flights to two facilities, and confirmed stabbings at five — including an attack on a high-ranking Blood leader at Hays State Prison during an official inspection, in front of the warden ("Blood on Blood," GPS, April 2026). On May 21, 2026, gang violence again forced a statewide lockdown, with deaths reported at Ware and Augusta and hospitalizations at Dooly (nine) and Central (six). At Ware State Prison — designed for roughly 500, holding about 1,447 — coroner and GBI records obtained by GPS through an open-records request document at least 19 prisoner-on-prisoner homicides since 2020, alongside twelve suicides and eight fatal overdoses. The June 2024 torture of Christian Krauch at Macon State Prison — three weeks hidden under a bunk while approximately 168 mandatory headcounts were submitted showing all inmates accounted for — shows what "documentation without supervision" means in practice ("Three Weeks Under a Bunk," GPS, February 2026).

Here is the legislative fact the state would prefer no one notice: stopping this never required a new law. O.C.G.A. § 42-5-52, tracing to 1897, commands that the department "shall provide for the classification and separation of inmates with respect to age, first offenders, habitual criminals and incorrigibles…" — a duty, not a courtesy ("On the Books Since 1897," GPS, May 2026). Texas, Arizona, and California built housing-based separation, intelligence-driven classification, and structured exit programs decades ago, with documented results. Georgia's refusal to enforce its own statute is the single cheapest reform available: separation costs nothing; the state keeps choosing the bodies ("Separate the Gangs. It Costs Nothing.," GPS, May 2026).

8. Vision 2027: The Post-Conviction Reform Package

Vision 2027 is GPS's post-conviction justice reform framework — the synthesis of more than a dozen research collections covering the economics of incarceration, the legal architecture of accountability, comparative reforms from other states, and the specific statutory levers that exist in Georgia law. For committee staff, its components translate into distinct bill files:

• Population management with a court-tested record. The framework assembles court-tested population-reduction mechanisms and costed oversight models from other states, targeted at the people for whom incarceration is most expensive and least justifiable: the 5,753 incarcerated people aged 60+, the terminally ill, and the backlog of 2,542 state inmates sitting in county jails on county budgets.
• Geriatric and medical release expansion. The recent mortality record — deaths at 85, 77, 66, 65, and 62 within a single month — quantifies the alternative: Georgia currently provides end-of-life care at correctional prices inside facilities under a federal deliberate-indifference finding.
• Sentencing and review modernization. The package catalogs statutory levers unique to Georgia that have never been fully used.

Wrongful-conviction landscape — and an honest data note. Post-conviction integrity is where the evidence base is thinnest, and members should know exactly what is and is not documented. On the documented side: in July 2026, an Atlanta Journal-Constitution report covered a father granted a new trial after more than two decades in prison in his son's death — a reminder that Georgia's post-conviction machinery does occasionally surface catastrophic error. On the undocumented side: GPS does not yet have public data quantifying error rates, backlog, or audit findings for the GBI crime laboratory system, and no member should cite figures that do not exist. What GPS can document is the adjacent integrity problem: death investigations inside prisons where the medical examiner's manner-of-death determination rested on GDC's own incident reports, and autopsy-to-agency-record discrepancies GPS has flagged at Telfair. If the state's forensic pipeline cannot be trusted to establish how a person died in custody, the same pipeline's role in establishing guilt deserves committee attention. Entry points — post-conviction discovery standards, evidence preservation mandates, and funding for conviction-integrity review — are available without prejudging the data.

9. End the Warehouse: Rehabilitation ROI and Comparable-State Evidence

GPS's End the Warehouse campaign assembles the evidentiary case that Georgia's crisis is a design choice, not a management problem — and that the alternative is costed, tested, and cheaper:

• The returns are measured, not hoped for. RAND: correctional education cuts recidivism 43 percent and returns $4–5 per $1. GDC's own vocational completers recidivate at about half the general rate. The FY2027 budget invests roughly $26 per person per year in rehabilitation against a system whose failures produced a $50.6-million settlement floor and a federal findings letter.
• The model has a name and American examples. The normalization principle — life inside should resemble life outside as closely as security allows, because the punishment is the loss of liberty and nothing beyond it — is operating in Scandinavia and is now being built on American soil; California's emerging model redesigns prisons around small-unit housing, education, and work ("Normalization: The Principle That Changes Everything," GPS, November 2025; "Reopen the Doors," GPS, June 2026).
• The vacuum is the policy failure. GPS's series documents the mechanism: strip a prison of work, school, and family contact, and the gang becomes — in the words of the series title — "The Only Family Left," supplying the purpose and protection the state withdrew. GDC's Securus commission structure monetizes the severed family line at 59.6 cents on the dollar (family communication analysis).
• District lift. Every point of recidivism reduction is measured in fewer victims in members' districts; every geriatric release is medical cost shifted off the most expensive care setting in state government; every county-jail backlog bed returned is a county budget relieved. Rehabilitation is not a programmatic nicety layered on the budget — it is the only line item in it with a documented positive return.

10. Policy Recommendations and Bill-Language Entry Points

Each recommendation is grounded in the record above; each entry point identifies where language attaches.

1. Enforce the 1897 separation duty. Entry point: Amend O.C.G.A. § 42-5-52 to add (a) a statutory definition of security-threat-group separation housing, (b) an annual GDC report to the judiciary committees certifying rival-faction housing assignments by facility, and (c) a structured gang-renouncement pathway. The statute already says "shall"; the amendment adds the reporting that makes "shall" enforceable. Cost: negligible. (Record basis: 315 gangs, 31% validation rate, three statewide lockdowns in five months.)

2. Death-in-custody transparency. Entry point: New reporting mandate in Title 42 requiring public disclosure of every in-custody death — name, facility, date, and preliminary manner — within 30 days, with final cause-of-death reporting within 90; establish an independent fatality-review panel outside GDC. (Record basis: at least 1,859 GPS-tracked deaths; 1,044 unknown/pending; two-month reporting lag; autopsy-record discrepancies at Telfair; "The Empty Column.")

3. Truth in vacancy numbers. Entry point: Appropriations rider requiring facility-level certification of security-post vacancy rates, excluding transitional-center and administrative staffing from security averages, before release of supplemental security funds. (Record basis: Zirkle's on-record "skewed" testimony; six officers on duty at Washington SP per GPS's open-records production; Guidehouse "emergency mode.")

4. A food floor with existing inspection data. Entry point: Appropriations rider phasing the per-meal floor from $0.54 toward the ACA benchmark over three fiscal years, paired with a mandate that DPH transmit all correctional-kitchen inspection scores to a centralized public portal. The inspection apparatus already exists under § 26-2-370; only aggregation is new. (Record basis: Johnson 64, Pulaski 67, Smith 68; 29-point seven-day score swings.)

5. Grievance-process minimum standards. Entry point: Statutory minimums overriding SOP 227.02's exclusion architecture — tolling of the ten-day filing window, acceptance of multi-issue grievances, extension of the seven-day appeal window, and anti-retaliation protection for filers. Under the PLRA, Georgia's grievance machine is the gatekeeper to federal court; a process designed to produce procedural default produces both suppressed claims and the 61 retaliation-linked events GPS has catalogued (retaliation analysis; "The Grievance Machine," GPS, July 2026).

6. PREA investigation integrity. Entry point: Independent state-level review of a sampling of PREA investigation files annually, with findings reported to the judiciary committees; audit-standards legislation ending the practice of GDC contracting its own auditors without external validation. (Record basis: 273 audits, 100% compliance vs. 388 files, 0 compliant, per GDC's own consultants; 3.5% substantiation across 15,542 allegations.)

7. Restrictive-housing duration caps. Entry point: Statutory cap on continuous restrictive-housing placement with mandatory periodic review hearings and public reporting of long-term stays. (Record basis: 26% of SMU population held 5+ years as of 2017; federal litigation already forced partial reform; half of suicides linked to isolation per GPS's mental health analysis.)

8. Geriatric and medical release. Entry point: Expand eligibility criteria and streamline processing for compassionate/medical release; require annual reporting on applications, grants, and deaths of approved-but-unreleased individuals. (Record basis: 5,753 incarcerated people aged 60+; recent deaths at 85, 77, 66, 65, 62; health-program costs up $37.3 million in two cycles.)

9. Inmate Welfare Fund and communications transparency. Entry point: Cap commissions on inmate communications contracts and require annual public accounting of Inmate Welfare Fund balances and expenditures. (Record basis: 59.6% Securus commission, ~$8M/year, $325K monthly floor, per contract documents.)

10. Independent oversight with subpoena power. Entry point: Create a correctional ombudsman or legislative oversight office with document-subpoena and facility-access authority. (Record basis: DOJ's "unnecessarily contentious" records fight requiring subpoena enforcement; pre-inspection cleanups; federal contempt for falsified records; the Senate Study Committee on DOC Facilities' own documentation of the crisis.) GDC publicly rejected DOJ's findings as a "fundamental misunderstanding"; an agency that grades itself needs a grader.

11. District-Relevant Facilities by Region

GPS maintains a public data page for every tracked facility — mortality records, claims, inspection scores, and issue links — at gps.press/facilities-data/ (machine-readable). Selected facilities by region:

Northwest Georgia

Hays State Prison (site of the April 2026 inspection-day stabbing) · Walker State Prison · Walker County Prison · Floyd County Prison · Chattooga County Prison · Gilmer County Prison

Metro Atlanta

Phillips State Prison · Atlanta Transitional Center · Metro Reentry Facility · Clayton Transitional Center · Fulton County Prison · Henry County Prison · Spalding County Prison · Gwinnett County Prison · Coweta County Prison · Carroll County Prison · Paulding PDC

Northeast Georgia

Arrendale State Prison · Whitworth Women's Facility · Athens/Clarke County Prison · Hall County Prison · Walton County Prison · Wilkes County Prison · Hart County Prison

East Central Georgia (Augusta–Sandersville corridor)

Augusta State Medical Prison (repeated recent deaths) · Hancock State Prison (70–80% vacancy testimony) · Washington State Prison (four homicides Jan. 11, 2026; six officers on duty) · Johnson State Prison (lowest kitchen score on record, 64) · Richmond County Prison · Jefferson County Prison · Emanuel Women's Facility · Screven County Prison · Bulloch County Prison · Jenkins Facility

Middle Georgia

Central State Prison (multiple recent family-attestation assault allegations) · Baldwin State Prison · Macon State Prison (Krauch torture case; 70–80% vacancy testimony) · Burruss CTC · Pulaski State Prison (failing kitchen score; recent deaths) · Dodge State Prison · Putnam State Prison · Georgia Diagnostic and Classification Prison and the Special Management Unit (long-term solitary) · Macon Transitional Center · Upson County Prison

Southwest Georgia

Lee State Prison · Calhoun State Prison (70–80% vacancy testimony) · Autry State Prison (70–80% vacancy testimony; defunded $150K peer-led reentry pilot) · Albany Transitional Center · Dougherty County Prison · Sumter County Prison · Terrell County Prison · Mitchell County Prison · Colquitt County Prison · Decatur County Prison · Grady County Prison · Thomas County Prison · Miller County Prison · Bainbridge Probation RSAT

Southeast & Coastal Georgia

Ware State Prison (at least 19 homicides since 2020 per coroner records) · Smith State Prison (former warden indicted May 2026) · Rogers State Prison · Georgia State Prison · Coastal State Prison · Chatham State Prison · Effingham County Prison · Long Unit · Montgomery State Prison · Appling ITF · Bacon Transitional Center

South Georgia (Valdosta–Tifton corridor)

Valdosta State Prison (80% officer vacancy; repeated recent deaths) · Valdosta Transitional Center · Lowndes Unit · Dooly State Prison (nine hospitalized May 21, 2026) · Wilcox State Prison · Coffee Correctional Facility (private) · Wheeler Correctional Facility (private) · Telfair State Prison (70–80% vacancy testimony; autopsy-record discrepancies) · Tift County Prison · Turner County Prison · Cook County Prison

West Central Georgia

Rutledge State Prison · Columbus Transitional Center · Muscogee County Prison · Harris County Prison · Meriwether County Prison

Take the Next Step

The federal findings are written. The settlement ledger is compiled. The mortality archive is public, and the separation statute has been on the books since 1897. What has been missing is a legislature willing to read its own state's documents.

GPS offers members and staff three concrete actions:

1. Schedule a confidential staff briefing. GPS will walk committee analysts through the settlement ledger, the facility-level vacancy evidence, the Ware coroner record set, and the line-item budget analysis — with documentation. Request through the GPS Intelligence System.
2. Request the data extracts. Facility-level mortality, claims, and inspection data for any district are available in machine-readable form at gps.press/facilities-data/; the full mortality archive is at gps.press/georgia-prisons-deaths archive. Deep-dive briefings on every topic in this memo — from legal settlements to staff misconduct to legal access — are published and citable.
3. Sponsor the bills. The entry points in Section 10 are drafted to attach to existing statutes and appropriations riders: enforce § 42-5-52, mandate death reporting within 30 days, certify facility-level security vacancy before releasing supplemental funds, aggregate the DPH kitchen scores the state already collects, and create oversight with subpoena power. Georgia is already paying for the crisis — $1.79 billion a year, $50.6 million in settlements and counting, at least 1,859 dead since 2020. The only remaining question is whether the General Assembly keeps paying for failure or starts paying for what works.