First-Line Supervisors of Correctional Officers

33-1011.00
Median wage $77,970/yr53,380 employed (US)Rank #876 of 923 scored · top 95% by substitution

Directly supervise and coordinate activities of correctional officers and jailers.

Sub-scores

0–100 · band = confidence interval from rater disagreement

Substitution10
Exposure11
Augmentation35

Substitution — the headline: capability discounted by cost, barriers and adoption.

Exposure — technical capability alone, regardless of whether anyone deploys it.

Augmentation — how much AI assists without replacing. High here + moderate substitution = a changing job, not a disappearing one.

Tasks on the substitution scale

23 rated tasks, binned by substitution score.

Position among all scored occupations

Distribution of 923 occupation scores; the marker is this occupation.

Tasks with substitution ≥ 70

0%

Run 1.0.0-draft.1 · computed 2026-08-05 · rater panel: claude-sonnet-5, claude-haiku-4-5-20251001 · intervals span rater disagreement.

Why this score

The five weighted dimensions of the composite, averaged across this occupation's tasks (importance-weighted, panel mean). Exact weights and formulas: /api/v1/methodology.

Task automatabilityw 35%12

panel mean rating 1.5/5 → substitution pressure 12/100

Technical feasibility todayw 20%11

panel mean rating 1.4/5 → substitution pressure 11/100

Cost vs. human wagew 15%13

panel mean rating 1.5/5 → substitution pressure 13/100

Adoption barriersw 20%inverted — strong barriers lower the score10

panel mean rating 4.6/5 (barrier strength) → substitution pressure 10/100

Sector adoption velocityw 10%3

panel mean rating 1.1/5 → substitution pressure 3/100

Task breakdown (23 tasks)

Substitution pressure per task, weighted by O*NET importance in the composite. Expand a task for the full rater audit trail — every rating, every model, every rationale.

Set up employee work schedules.

57

CI 48–67 · exposure 62 · augmentation 75 · importance 4.2/5 · click for rater detail

Panel ratings per dimension for this task
DimensionModelRatingRationale
Sector adoption velocityclaude-haiku-4-5-202510013/5Public-sector correctional systems have adopted scheduling software, but adoption is uneven and slower than in private industry; many facilities still rely on manual or semi-manual approaches.
Sector adoption velocityclaude-sonnet-52/5Corrections is a slow-moving, heavily unionized government sector with limited digitization and low AI adoption compared to information or finance sectors.
Augmentation potentialclaude-haiku-4-5-202510014/5Scheduling tools consistently augment supervisor productivity by automating routine assignment and constraint checks, freeing the supervisor to focus on adjustments, exceptions, and strategic staffing decisions.
Augmentation potentialclaude-sonnet-54/5AI-assisted scheduling tools can significantly speed up draft creation, flag conflicts, and optimize coverage, meaningfully boosting a supervisor's productivity while they retain final approval.
Task automatabilityclaude-haiku-4-5-202510014/5Scheduling algorithms can handle the majority of shift-assignment logic (staffing levels, rotation patterns, constraint satisfaction), achieving substantial time savings and quality parity with manual scheduling. However, edge cases—conflict resolution, special accommodations, union rules—may still require human judgment.
Task automatabilityclaude-sonnet-53/5Scheduling software with constraint optimization can generate valid staff rosters covering shift rules, coverage minimums, and seniority, automating much of the drafting work, but corrections settings involve frequent overrides for security incidents, union rules, and last-minute absences that require human judgment.
Adoption barriersclaude-haiku-4-5-202510013/5Union agreements and collective bargaining often constrain scheduling changes, and some jurisdictions may require supervisory sign-off on schedules. These create moderate friction but do not legally mandate a human perform the automation itself.
Adoption barriersclaude-sonnet-53/5No licensing requirement mandates a human create schedules, but union contracts, staffing regulations for correctional facilities, and security-sensitive coverage requirements create real organizational friction against pure automation.
Cost vs. human wageclaude-haiku-4-5-202510014/5Scheduling software (subscription or one-time license) spreads across many employees and pay periods, making per-task cost very low compared to a supervisor's loaded wage for manual scheduling work.
Cost vs. human wageclaude-sonnet-53/5Scheduling software licenses are inexpensive relative to a supervisor's time spent, but the supervisor still needs to review, approve, and adjust schedules, so total cost savings are moderate rather than order-of-magnitude.
Technical feasibility todayclaude-haiku-4-5-202510014/5Scheduling software is deployed in many correctional facilities and other organizations; off-the-shelf tools reliably generate compliant shift patterns. Products exist and function well in production, though some customization for facility-specific rules is often needed.
Technical feasibility todayclaude-sonnet-53/5Workforce scheduling software is widely deployed in corrections and similar shift-based industries, but it typically requires human review and manual adjustment rather than fully autonomous operation.

Complete administrative paperwork or supervise the preparation or maintenance of records, forms, or reports.

36

CI 25–48 · exposure 38 · augmentation 63 · importance 4.3/5 · click for rater detail

Panel ratings per dimension for this task
DimensionModelRatingRationale
Sector adoption velocityclaude-haiku-4-5-202510012/5Correctional agencies are among the slowest adopters of digital technology, with many still using paper-based or decades-old legacy systems. Budget constraints, resistance to change, and the specialized nature of corrections IT infrastructure mean AI-driven administrative automation is rare in actual deployment despite theoretical feasibility.
Sector adoption velocityclaude-sonnet-52/5Government and correctional institutions are typically slow adopters of AI due to security, legacy systems, and public-sector procurement constraints, resulting in low observed deployment of AI in this domain.
Augmentation potentialclaude-haiku-4-5-202510013/5AI can meaningfully assist supervisors by auto-populating routine fields, flagging missing information, and generating draft reports from incident data—raising productivity for those parts of the workflow. However, the supervisory review and judgment component remains central, so augmentation is moderate rather than transformative.
Augmentation potentialclaude-sonnet-54/5AI drafting, summarization, and template-filling tools can meaningfully speed up completion and review of administrative paperwork while the supervisor remains responsible for accuracy and sign-off.
Task automatabilityclaude-haiku-4-5-202510012/5Administrative paperwork in corrections is heavily standardized and rule-based, making parts of it suitable for automation (form filling, report generation from structured data). However, supervisory review, verification against institutional policies, and judgment calls about what constitutes appropriate documentation require human oversight that cannot be fully automated while maintaining accountability—limiting time savings to roughly 20–40%.
Task automatabilityclaude-sonnet-53/5Drafting standard reports, filling templated forms, and summarizing records can be substantially automated with document-generation tools, but supervisory review, data verification, and institution-specific record-keeping still require human judgment and accountability.
Adoption barriersclaude-haiku-4-5-202510014/5Corrections operates under strict regulatory requirements (DOJ standards, state compliance rules, litigation discovery obligations) where records are legally consequential. Supervisors are typically accountable for record accuracy and completeness, creating a legal and liability barrier that prevents full substitution without human sign-off—organizational and regulatory friction is high.
Adoption barriersclaude-sonnet-53/5Correctional records often have legal evidentiary value and require accountability chains and supervisor sign-off, creating moderate barriers even though the paperwork itself isn't a licensed-only task.
Cost vs. human wageclaude-haiku-4-5-202510012/5AI inference and integration costs are modest, but the oversight burden is substantial: supervisors must still review, validate, and correct AI-generated documents to meet legal and safety standards. The all-in cost per completed record approaches human wage costs when oversight is factored in, particularly given the liability sensitivity of correctional records.
Cost vs. human wageclaude-sonnet-53/5Software tools for drafting and organizing paperwork are cheap relative to a supervisor's wage, but integration with secure correctional records systems and required oversight narrows the savings.
Technical feasibility todayclaude-haiku-4-5-202510013/5General document automation and form-filling tools exist and are deployed in some correctional facilities, but most corrections agencies still rely on legacy systems with limited AI integration. Products can handle routine data entry and basic report generation, but the integration with facility-specific record systems and compliance requirements remains inconsistent and error-prone.
Technical feasibility todayclaude-sonnet-52/5General office AI (e.g., document assistants, OCR, workflow software) exists but corrections-specific record systems (incident reports, custody logs) are rarely fully AI-driven in production; most facilities still rely on manual entry and human-supervised review.

Review offender information to identify issues that require special attention.

23

CI 23–23 · exposure 25 · augmentation 50 · importance 4.0/5 · click for rater detail

Panel ratings per dimension for this task
DimensionModelRatingRationale
Sector adoption velocityclaude-haiku-4-5-202510011/5Correctional agencies are historically slow to adopt AI; most facilities still rely on manual record review and established protocols. Digital transformation in corrections lags information and finance sectors, and budget constraints and staff resistance limit pilot adoption.
Sector adoption velocityclaude-sonnet-51/5Corrections is a low-digitization, government-run sector with limited AI tool deployment and slow bureaucratic adoption cycles.
Augmentation potentialclaude-haiku-4-5-202510013/5AI can assist by flagging anomalies, summarizing case histories, and surfacing relevant prior incidents, helping supervisors focus human attention more efficiently. However, the core judgment remains human-centered, so augmentation is moderate rather than transformative.
Augmentation potentialclaude-sonnet-53/5AI can help summarize case files, highlight anomalies, or flag risk indicators, meaningfully aiding a supervisor's review process even though final judgment remains human-driven.
Task automatabilityclaude-haiku-4-5-202510012/5AI can extract and flag information from structured records (names, charges, prior incidents) but identifying 'issues that require special attention' requires contextual judgment about custody risk, behavioral patterns, and institutional priorities that vary by facility and situation. Current systems cannot reliably perform the full interpretive task end-to-end.
Task automatabilityclaude-sonnet-52/5AI can flag risk factors or summarize offender records, but interpreting context, institutional history, and safety implications requires human judgment that current systems cannot fully replicate end-to-end.
Adoption barriersclaude-haiku-4-5-202510014/5Correctional facilities operate under strict regulatory oversight and liability frameworks; supervisory decisions directly affect custody and safety, creating legal exposure if AI errors lead to incidents. Human accountability and sign-off are expected norms in this sector, and automation faces organizational and legal friction.
Adoption barriersclaude-sonnet-54/5Correctional supervision involves institutional liability, security/safety regulations, and legal accountability that generally require a certified human officer to make final judgment calls on offender risk.
Cost vs. human wageclaude-haiku-4-5-202510012/5Integration of AI into legacy correctional information systems is costly and requires custom setup; ongoing human oversight to validate alerts remains essential. Total cost (inference, integration, oversight) approaches or exceeds the cost of human review, given the consequences of missed security risks.
Cost vs. human wageclaude-sonnet-52/5While AI text analysis is cheap, the sensitive nature of this task requires substantial human oversight and verification, keeping effective all-in costs closer to human-comparable levels.
Technical feasibility todayclaude-haiku-4-5-202510012/5While NLP tools can surface keywords and anomalies in offender databases, no deployed product reliably performs the supervisory judgment of prioritizing which issues warrant intervention in a correctional setting. Prototype systems exist but lack the contextual accuracy and liability clearance needed for production use.
Technical feasibility todayclaude-sonnet-52/5Some correctional/case management systems use flagging algorithms or risk-assessment tools, but comprehensive automated review of offender files for 'special attention' issues is not a mature, widely deployed product function.

Conduct evaluations of employees' performance.

23

CI 23–23 · exposure 25 · augmentation 38 · importance 3.8/5 · click for rater detail

Panel ratings per dimension for this task
DimensionModelRatingRationale
Sector adoption velocityclaude-haiku-4-5-202510011/5Correctional institutions are slow-moving, unionized public sector organizations with limited digitization and strong resistance to automation in authority-sensitive functions. Adoption of AI for performance evaluation in this sector is minimal and lagging far behind private or tech-forward sectors.
Sector adoption velocityclaude-sonnet-51/5Corrections is a public-sector, physically-oriented field with low digitization and slow AI adoption for personnel management functions.
Augmentation potentialclaude-haiku-4-5-202510012/5AI could assist with scheduling evaluation reminders, summarizing documented incidents, or flagging policy violations, but the core task—forming judgments about officer conduct, reliability, and fit—remains supervisory and context-dependent, limiting meaningful augmentation.
Augmentation potentialclaude-sonnet-53/5AI can help supervisors organize notes, incident records, and draft evaluation language, providing moderate assistance while the supervisor retains judgment and final assessment.
Task automatabilityclaude-haiku-4-5-202510012/5Performance evaluations require contextual judgment about subjective behaviors, interpersonal dynamics, and disciplinary assessment that current AI cannot reliably execute end-to-end. While AI could assist with data aggregation and documentation, the legal and human accountability requirements in correctional settings demand human supervisor discretion and sign-off.
Task automatabilityclaude-sonnet-52/5Performance evaluation of correctional officers requires judgment about observed conduct, teamwork, and situational handling that AI cannot directly observe or assess reliably, though it can help draft or aggregate input.rief.5
Adoption barriersclaude-haiku-4-5-202510014/5Correctional settings are heavily regulated, and first-line supervisors have legal and contractual authority to evaluate subordinates; evaluation outcomes affect discipline, termination, and union grievances. Liability asymmetry is high—an AI error in a performance evaluation could trigger litigation or administrative backlash.
Adoption barriersclaude-sonnet-54/5Performance evaluations affecting discipline, promotion, or termination in a correctional/government setting typically require documented human judgment and supervisory sign-off, with union and civil-service protections limiting automation.
Cost vs. human wageclaude-haiku-4-5-202510012/5Integration of AI evaluation tools, combined with mandatory human review and liability oversight in a correctional context, would approach or exceed the cost of a supervisor conducting the evaluation directly, especially given relatively small supervisory spans in prisons.
Cost vs. human wageclaude-sonnet-52/5AI drafting tools are cheap but still require the supervisor's full observational input and judgment, so overall cost savings versus the human process are modest.
Technical feasibility todayclaude-haiku-4-5-202510012/5No deployed AI systems perform independent correctional employee evaluations in production. Tools exist for generic HR evaluation support, but none are calibrated for the behavioral and compliance nuances of correctional officer performance in real carceral environments.
Technical feasibility todayclaude-sonnet-52/5Some HR software offers AI-assisted performance review drafting and summarization, but no deployed product independently evaluates correctional staff performance based on job-specific conduct observations.

Examine incoming or outgoing mail to ensure conformance with regulations.

20

CI 18–23 · exposure 25 · augmentation 50 · importance 4.2/5 · click for rater detail

Panel ratings per dimension for this task
DimensionModelRatingRationale
Sector adoption velocityclaude-haiku-4-5-202510011/5Correctional systems are conservative, resource-constrained, and operate under high regulatory friction. Adoption of AI mail screening remains minimal; most facilities still rely on manual inspection, and organizational inertia is substantial.
Sector adoption velocityclaude-sonnet-51/5Corrections is a slow-adopting, physically-oriented government sector with limited digitization and cautious technology rollout given security and liability concerns.
Augmentation potentialclaude-haiku-4-5-202510013/5AI-assisted imaging and pattern flagging can help supervisors prioritize and speed review of suspect items, but the supervisory judgment and final sign-off remain human-centered. Moderate productivity gain is plausible with well-integrated tools.
Augmentation potentialclaude-sonnet-53/5AI-assisted scanning, OCR, and image analysis can help supervisors triage or flag mail faster, providing useful augmentation even though human judgment and physical inspection remain central.
Task automatabilityclaude-haiku-4-5-202510012/5Limited automation is possible for detecting prohibited items (drugs, weapons) via imaging/scanning, but the task requires nuanced judgment of regulatory conformance, intent assessment, and contextual discretion that current AI systems cannot reliably perform end-to-end. Meaningful human oversight remains mandatory.
Task automatabilityclaude-sonnet-52/5AI can flag suspicious content, keywords, or contraband-related patterns in text/images, but final judgment on regulatory conformance, contraband detection, and security risk in correctional mail requires human review and physical inspection.4 The task also involves physical handling not automatable by software alone.
Adoption barriersclaude-haiku-4-5-202510015/5Correctional mail inspection is legally mandated and carried out under strict regulatory authority (state/federal DOC rules). A licensed supervisory officer must remain responsible for compliance decisions; liability and legal custody duties create hard barriers to full automation.
Adoption barriersclaude-sonnet-54/5Security-sensitive correctional operations require accountable staff for chain-of-custody, legal compliance, and contraband interdiction, creating strong institutional and regulatory barriers against full automation.
Cost vs. human wageclaude-haiku-4-5-202510012/5Advanced imaging and AI screening systems carry significant capital and integration costs, and still require supervisory oversight. The all-in cost per mail item checked likely exceeds the marginal cost of human inspection in most correctional settings.
Cost vs. human wageclaude-sonnet-52/5Physical mail handling, scanning equipment, and human oversight remain necessary; AI tools add cost on top of required human review rather than replacing the labor outright, keeping costs comparable to human-only baseline.
Technical feasibility todayclaude-haiku-4-5-202510012/5AI-assisted mail screening (imaging, OCR) exists in some facilities, but no deployed system autonomously performs full regulatory conformance verification at production scale. Products are narrowly scoped and require substantial human review.
Technical feasibility todayclaude-sonnet-52/5Some correctional facilities use scanning/imaging software and AI-assisted contraband detection tools, but these are narrow-scope aids, not full end-to-end mail conformance review systems deployed reliably at scale.

Convey correctional officers' or inmates' complaints to superiors.

20

CI 18–23 · exposure 20 · augmentation 50 · importance 4.0/5 · click for rater detail

Panel ratings per dimension for this task
DimensionModelRatingRationale
Sector adoption velocityclaude-haiku-4-5-202510011/5Correctional facilities are among the slowest-adopting sectors for AI automation, characterized by legacy systems, limited digitization, and strong institutional resistance to removing human judgment from sensitive processes. Pilot programs are rare; production deployment of autonomous complaint systems is negligible.
Sector adoption velocityclaude-sonnet-51/5Corrections is a low-digitization, physically embedded government sector with minimal AI adoption in supervisory communication workflows.
Augmentation potentialclaude-haiku-4-5-202510013/5AI could usefully assist by auto-summarizing or categorizing incoming complaints and flagging urgent cases, raising a supervisor's efficiency in processing high complaint volume. However, the assistant role is modest—human judgment remains essential for final assessment and escalation decisions.
Augmentation potentialclaude-sonnet-53/5AI can help draft, log, categorize, and summarize complaints for faster relay to superiors, but the supervisor remains essential for judgment and interpersonal handling.
Task automatabilityclaude-haiku-4-5-202510012/5While AI could help draft complaint summaries or route messages, the task requires interpreting nuanced human complaints, assessing credibility, and determining appropriate escalation—judgment calls that remain substantially human-dependent. Current systems cannot reliably capture context-sensitive complaint nuance or make defensible escalation decisions in a correctional environment without human oversight, falling short of 50% time-saving parity.
Task automatabilityclaude-sonnet-52/5Conveying complaints involves receiving, interpreting, contextualizing, and relaying sensitive human information within a chain of command; AI could draft summaries but cannot independently judge escalation priority or handle interpersonal nuance reliably.'
Adoption barriersclaude-haiku-4-5-202510014/5Correctional settings are heavily regulated, and complaint handling (especially from inmates) may have statutory or liability requirements that a human supervisor must formally sign off on. Organizational and legal culture in corrections strongly favors documented human accountability in complaint chains, creating material friction to full automation.
Adoption barriersclaude-sonnet-54/5Correctional oversight involves legal accountability, chain-of-command requirements, and inmate rights protections, meaning a human supervisor typically must handle and be responsible for conveying complaints.
Cost vs. human wageclaude-haiku-4-5-202510012/5A supervisor's loaded wage is ~$70–90K annually; AI for complaint routing and drafting (including oversight and integration) would cost several thousand dollars monthly. When amortized against one supervisor's time, the cost ratio does not clearly favor automation, particularly given error-correction overhead.
Cost vs. human wageclaude-sonnet-52/5While transcription or summarization tools are cheap, the human judgment, trust-building, and accountability required for this task keep a supervisor's involvement necessary, limiting cost savings.
Technical feasibility todayclaude-haiku-4-5-202510012/5No deployed product reliably performs complaint intake, assessment, and escalation in correctional facilities. AI systems exist for basic message routing or summarization in some organizations, but none demonstrably handle the sensitive context and interpersonal judgment required in a correctional setting at production scale.
Technical feasibility todayclaude-sonnet-51/5No deployed correctional-facility product autonomously receives and escalates officer/inmate complaints to superiors; this remains a human supervisory function with legal and safety implications.

Develop work or security procedures.

18

CI 13–23 · exposure 20 · augmentation 50 · importance 4.0/5 · click for rater detail

Panel ratings per dimension for this task
DimensionModelRatingRationale
Sector adoption velocityclaude-haiku-4-5-202510011/5Correctional facilities are organizationally conservative, heavily regulated, and risk-averse regarding security matters. Adoption of AI for procedure development in this sector is minimal; procedures remain human-driven due to legal accountability and specialized domain requirements.
Sector adoption velocityclaude-sonnet-51/5Corrections is a low-digitization, government-run sector with minimal AI adoption for core security policy work, lagging far behind information/finance sectors.
Augmentation potentialclaude-haiku-4-5-202510013/5AI could assist supervisors by drafting procedure templates, organizing regulatory requirements, or identifying inconsistencies in existing protocols, but final development and authorization must remain with qualified human supervisors due to security and legal implications.
Augmentation potentialclaude-sonnet-53/5AI can help draft initial procedure language, summarize best practices, or check documents against regulations, offering moderate assistance while the supervisor retains final judgment.
Task automatabilityclaude-haiku-4-5-202510012/5Developing security procedures requires domain expertise, legal compliance knowledge, and contextual judgment about facility-specific risks that current AI systems cannot reliably produce end-to-end. While AI can draft templates or assist with documentation, the core task of designing effective security protocols demands human judgment, regulatory knowledge, and accountability that AI cannot assume.
Task automatabilityclaude-sonnet-52/5Drafting procedure text can be AI-assisted, but developing effective security procedures requires site-specific risk assessment, knowledge of inmate behavior patterns, and legal compliance judgment that current AI cannot reliably synthesize end-to-end.
Adoption barriersclaude-haiku-4-5-202510014/5Strong legal and regulatory barriers exist: correctional security procedures must comply with state and federal regulations, constitutional standards, and agency policies. Liability exposure is high, and courts expect procedures to be developed by qualified personnel; automation faces significant legal and organizational resistance.
Adoption barriersclaude-sonnet-55/5Correctional security procedures typically require sign-off by certified corrections supervisors and compliance with state/federal regulations, plus high liability if procedures fail, making this a hard-barrier task.
Cost vs. human wageclaude-haiku-4-5-202510012/5The cost of AI-generated security procedures that require extensive human review, legal vetting, and revision is comparable to or potentially exceeds the cost of expert human development, especially when accounting for the liability and oversight burden.
Cost vs. human wageclaude-sonnet-52/5While AI drafting tools are cheap, the human oversight, security vetting, and legal review needed for correctional procedures still dominate the cost, keeping overall savings modest.
Technical feasibility todayclaude-haiku-4-5-202510012/5No deployed AI product reliably develops institutional security procedures at scale in production correctional environments. Research systems can generate generic procedures, but real-world deployment in correctional facilities involves legal, safety, and liability considerations that require human expertise and sign-off.
Technical feasibility todayclaude-sonnet-51/5No deployed product autonomously creates correctional facility security procedures; this remains a specialized institutional function requiring human expertise and approval chains.

Instruct employees or provide on-the-job training.

17

CI 9–25 · exposure 13 · augmentation 50 · importance 4.3/5 · click for rater detail

Panel ratings per dimension for this task
DimensionModelRatingRationale
Sector adoption velocityclaude-haiku-4-5-202510012/5Correctional facilities remain relatively resistant to automation adoption; they are geographically dispersed, heavily regulated, and culturally oriented toward hierarchical, in-person supervision. Digitization and AI adoption lag significantly behind finance or professional services.
Sector adoption velocityclaude-sonnet-51/5Corrections is a low-digitization, physically embedded government sector with slow AI adoption and no evidence of production-scale AI replacing supervisory training tasks.
Augmentation potentialclaude-haiku-4-5-202510013/5AI can usefully assist supervisors by drafting training modules, tracking completion records, and identifying performance gaps, but does not fundamentally transform supervisor productivity on the core task of direct instruction and behavioral coaching.
Augmentation potentialclaude-sonnet-53/5AI can help create training materials, quizzes, or track compliance/certification records, providing moderate assistance while the supervisor still delivers hands-on instruction.
Task automatabilityclaude-haiku-4-5-202510012/5While AI can generate training content and standardized instructional materials, the dynamic, individualized, and relationship-dependent nature of on-the-job training—especially in a correctional setting requiring behavioral feedback and adaptive coaching—cannot be fully automated. AI falls short of the 50% time-saving threshold for this inherently supervisory task.
Task automatabilityclaude-sonnet-51/5Instructing correctional officers on procedures, security protocols, and physical/behavioral tactics requires hands-on demonstration, live supervision, and situational judgment in a secure facility that AI cannot perform end-to-end today.
Adoption barriersclaude-haiku-4-5-202510014/5Correctional agencies typically require supervisors to certify staff competency and maintain accountability for officer performance and safety—functions that carry legal and liability weight. Regulatory frameworks and institutional practices strongly prefer human supervisors to sign off on critical training and performance assessments.
Adoption barriersclaude-sonnet-54/5Correctional training often involves legal/regulatory certification requirements, safety-critical procedures, and use-of-force protocols that require sign-off by an authorized supervisor, creating strong institutional and liability barriers to automation.
Cost vs. human wageclaude-haiku-4-5-202510012/5AI-assisted training tools (LMS, video platforms) reduce some administrative overhead but require significant setup, integration, and human oversight. The total cost of AI infrastructure plus necessary human supervision remains comparable to or higher than the direct cost of supervisor-led training.
Cost vs. human wageclaude-sonnet-52/5While AI-based e-learning modules are cheap, the core on-the-job instruction and hands-on coaching still require a paid human supervisor, so overall cost savings versus the human-only baseline are minimal.
Technical feasibility todayclaude-haiku-4-5-202510012/5No deployed products reliably perform full on-the-job training in production correctional environments. Some learning management systems and video-based training exist, but these address only static content delivery, not the real-time coaching and interpersonal instruction supervisors provide.
Technical feasibility todayclaude-sonnet-51/5No deployed AI product conducts on-the-job training or direct instruction of correctional staff in real facility settings; this remains a human supervisory function.

Conduct roll calls of correctional officers.

16

CI 9–23 · exposure 17 · augmentation 38 · importance 4.3/5 · click for rater detail

Panel ratings per dimension for this task
DimensionModelRatingRationale
Sector adoption velocityclaude-haiku-4-5-202510011/5Corrections is a publicly managed, traditionally low-digitization sector with strong union representation, regulatory constraints, and institutional resistance to automation in security-critical functions. Adoption of AI for supervisory roll calls remains extremely limited.
Sector adoption velocityclaude-sonnet-51/5Corrections is a slow-adopting, highly regulated, physically-grounded sector with minimal AI agent deployment for security-critical supervisory tasks.
Augmentation potentialclaude-haiku-4-5-202510012/5AI could assist with attendance data logging or shift scheduling, but the core task of conducting a roll call—presence verification, briefings, and accountability decisions—remains primarily supervisor-driven with limited potential for AI to meaningfully enhance productivity.
Augmentation potentialclaude-sonnet-53/5Digital rostering, scheduling, and attendance-tracking tools can streamline record-keeping and flag discrepancies, aiding the supervisor without replacing the human verification role.
Task automatabilityclaude-haiku-4-5-202510011/5Roll calls require real-time verification of officer presence, accountability, and immediate responsiveness to unexpected absences or incidents—tasks that demand human judgment and presence. Current AI cannot reliably perform the accountability, attendance reconciliation, and dynamic decision-making required to replace this task end-to-end.
Task automatabilityclaude-sonnet-52/5Basic attendance/roll-call logging could be handled by digital check-in systems, but conducting roll call also involves visual verification, brief inspection, and command presence that current AI cannot perform end-to-end in a secure facility.'
Adoption barriersclaude-haiku-4-5-202510015/5Correctional work is heavily regulated; supervisors have explicit legal and operational responsibility for officer accountability, facility security, and chain-of-custody documentation. Personnel attendance and briefing in correctional settings typically require a licensed human supervisor to conduct and sign off on roll calls.
Adoption barriersclaude-sonnet-54/5Correctional settings require sworn/certified supervisory personnel with legal authority to verify staff presence and readiness for security purposes, creating strong regulatory and institutional barriers to full automation.
Cost vs. human wageclaude-haiku-4-5-202510012/5Implementing AI attendance and verification systems requires integration with existing correctional management systems, substantial setup, and ongoing oversight. The cost likely approaches or exceeds the wage of a supervisor conducting routine roll calls.
Cost vs. human wageclaude-sonnet-52/5While automated check-in kiosks are cheap, they cannot replace the supervisory verification and authority component of roll call, so full task substitution would still require human oversight, limiting cost savings.
Technical feasibility todayclaude-haiku-4-5-202510012/5While AI-powered scheduling and attendance tracking systems exist in some facilities, no deployed product reliably automates the full supervisory roll call process including real-time confirmation, briefing, and personnel decisions. Most corrections facilities still rely on manual or semi-automated systems.
Technical feasibility todayclaude-sonnet-52/5Digital attendance and scheduling software exists in many workplaces, but no deployed AI product independently conducts correctional roll calls in production; this remains a human supervisory function with software as a minor aid.

Take, receive, or check periodic inmate counts.

9

CI 0–18 · exposure 13 · augmentation 25 · importance 4.8/5 · click for rater detail

Panel ratings per dimension for this task
DimensionModelRatingRationale
Sector adoption velocityclaude-haiku-4-5-202510011/5Correctional facilities represent a low-digitization, safety-critical, and regulatory-heavy sector with minimal AI adoption. The nature of custody and security accountability makes this a laggard domain for automation.
Sector adoption velocityclaude-sonnet-51/5Corrections is a highly physical, security-focused, historically low-digitization sector with strong institutional resistance to removing human verification from custody and security functions.'
Augmentation potentialclaude-haiku-4-5-202510011/5AI offers no meaningful assistance in the core task of physically accounting for and verifying inmate presence and identity. This is a straightforward procedural count that does not benefit from machine learning or decision support.
Augmentation potentialclaude-sonnet-53/5Electronic tracking, camera systems, and automated alerts can help supervisors cross-check and expedite counts, improving efficiency, but the officer must still physically verify and take responsibility for the count.'
Task automatabilityclaude-haiku-4-5-202510011/5Counting inmates requires physical presence, verification of identities, and real-time accountability in a secure environment. Current AI cannot autonomously perform these security-critical verification tasks, which depend on direct observation and manual confirmation by authorized personnel.
Task automatabilityclaude-sonnet-52/5Physical presence counting and verifying inmates requires visual/physical confirmation of bodies in cells, which current AI cannot perform end-to-end; sensor/camera-based systems can assist but not fully replace human verification for security-critical counts.'
Adoption barriersclaude-haiku-4-5-202510015/5Correctional facilities are heavily regulated environments where inmate accountability is a mandatory legal duty. Security protocols, chain-of-custody requirements, and federal/state corrections law typically mandate that a licensed/authorized human directly perform or certify counts.
Adoption barriersclaude-sonnet-55/5Inmate counts are a core security and legal accountability function; correctional regulations and institutional policy require certified, authorized human staff to conduct and verify counts, with strict chain-of-custody and legal liability for errors (escapes, safety failures).'
Cost vs. human wageclaude-haiku-4-5-202510011/5The loaded cost of a first-line supervisor performing periodic counts is substantially lower than the integration, hardware, and oversight infrastructure needed for AI-based count verification, with no proven alternative deployment model.
Cost vs. human wageclaude-sonnet-52/5Sensor and camera infrastructure requires significant capital investment, ongoing maintenance, and human oversight, so total cost is not dramatically cheaper than having correctional officers perform counts, especially given liability requirements.'
Technical feasibility todayclaude-haiku-4-5-202510011/5No deployed AI system performs inmate counting as an autonomous task in correctional facilities. This task inherently requires authorized human presence and legal custody documentation, making end-to-end automation infeasible today.
Technical feasibility todayclaude-sonnet-52/5Some prisons use electronic tracking (RFID, biometric) and camera systems to assist counts, but these are supplementary tools, not autonomous systems performing verified inmate counts reliably at scale without human confirmation.'

Rate behavior of inmates, promoting acceptable attitudes and behaviors to those with low ratings.

7

CI 0–14 · exposure 8 · augmentation 25 · importance 3.8/5 · click for rater detail

Panel ratings per dimension for this task
DimensionModelRatingRationale
Sector adoption velocityclaude-haiku-4-5-202510011/5Correctional facilities are low-digitization, traditional hierarchical organizations with entrenched human-centric models and strong institutional resistance to algorithmic decisions affecting inmate treatment and safety.
Sector adoption velocityclaude-sonnet-51/5Corrections is a low-digitization, physically embedded government sector with minimal AI deployment for direct inmate supervision and behavioral assessment.
Augmentation potentialclaude-haiku-4-5-202510012/5AI could assist with flagging behavioral trends or scheduling interventions, but current systems offer limited meaningful support for the core task of rating qualitative behavior and promoting attitude change through direct engagement.
Augmentation potentialclaude-sonnet-52/5AI could help track behavioral incident data or flag patterns for review, but it cannot meaningfully assist the actual interpersonal rating and behavior-shaping interactions with inmates.
Task automatabilityclaude-haiku-4-5-202510012/5Assessing and influencing inmate behavior requires contextual judgment, emotional intelligence, and nuanced interpersonal interaction that current AI cannot reliably perform end-to-end. While AI could support data analysis of behavioral patterns, the core task of rating behavior and promoting attitudinal change depends on human discretion and direct rapport.
Task automatabilityclaude-sonnet-51/5This requires in-person judgment, authority, and behavioral intervention with incarcerated individuals in a secure facility setting; no AI system can perform the interpersonal supervision or corrective action involved.'
Adoption barriersclaude-haiku-4-5-202510014/5Correctional facilities are highly regulated environments where accountability for inmate treatment and safety falls legally on human supervisors. Liability, duty of care, and regulatory requirements around inmate interactions create strong legal and organizational barriers to full automation.
Adoption barriersclaude-sonnet-55/5Corrections is a heavily regulated, security-sensitive government function requiring sworn/certified officers with legal authority over inmates, and safety and liability concerns strongly prohibit automation of behavioral rating and correction.
Cost vs. human wageclaude-haiku-4-5-202510011/5Human supervisors are already on-site performing this task; AI system deployment would represent additional cost with uncertain oversight savings, making AI more expensive than the status quo.
Cost vs. human wageclaude-sonnet-51/5There is no viable AI substitute providing this output, so no meaningful cost comparison favors AI; human supervisory staff are required.
Technical feasibility todayclaude-haiku-4-5-202510011/5No deployed product reliably performs behavioral assessment and attitude modification for incarcerated populations in production correctional settings. The task requires real-time judgment, safety considerations, and accountability that remain human-led.
Technical feasibility todayclaude-sonnet-51/5No deployed AI product performs inmate behavioral rating or corrective coaching; this remains entirely a human correctional function today.

Maintain knowledge of, comply with, and enforce all institutional policies, rules, procedures, and regulations.

4

CI 0–9 · exposure 5 · augmentation 38 · importance 4.7/5 · click for rater detail

Panel ratings per dimension for this task
DimensionModelRatingRationale
Sector adoption velocityclaude-haiku-4-5-202510011/5Correctional systems are traditionally slower to adopt automation due to union contracts, safety-critical liability concerns, and the requirement for human discretion and accountability in institutional governance. Adoption of AI for policy enforcement remains minimal.
Sector adoption velocityclaude-sonnet-51/5Corrections is a highly regulated, low-digitization public-sector environment with minimal AI agent deployment in operational enforcement roles.
Augmentation potentialclaude-haiku-4-5-202510013/5AI could usefully assist supervisors by maintaining searchable policy databases, flagging rule violations for review, and generating compliance reports, which would help supervisors stay current with institutional rules more efficiently while they retain final authority.
Augmentation potentialclaude-sonnet-52/5AI can help track policy updates, generate compliance checklists, or flag documentation gaps, but has limited direct impact on the interpersonal enforcement and judgment core to this task.
Task automatabilityclaude-haiku-4-5-202510011/5This task requires subjective judgment, institutional memory, and discretionary enforcement decisions that depend on context, precedent, and human interpretation. Current AI cannot reliably make enforcement decisions or ensure compliance with evolving institutional policies without human oversight.
Task automatabilityclaude-sonnet-51/5Enforcing policies and rules within a correctional facility requires physical presence, authority, judgment, and real-time human interaction that current AI cannot replicate or execute end-to-end.
Adoption barriersclaude-haiku-4-5-202510015/5Correctional institutions operate under strict legal and regulatory frameworks where a licensed supervisor must legally sign off on policy enforcement and compliance decisions. Liability for improper enforcement and the statutory requirement for human supervisory authority create hard barriers to substitution.
Adoption barriersclaude-sonnet-55/5This is a government-authorized supervisory role with legal accountability, use-of-force oversight, and staff safety implications that require a licensed, sworn human supervisor.
Cost vs. human wageclaude-haiku-4-5-202510012/5Integration of AI tools for policy tracking would require significant customization to institutional rule sets and workflow, and the cost savings would be modest given that human supervisors must still make final enforcement decisions and validate compliance determinations.
Cost vs. human wageclaude-sonnet-51/5There is no viable AI substitute performing this task, so cost comparison favors the human by default; any AI role would only supplement, not replace, at added cost.
Technical feasibility todayclaude-haiku-4-5-202510012/5While AI could assist with policy documentation retrieval and flagging potential violations, no deployed system reliably performs the full enforcement and compliance-monitoring function end-to-end. The task requires human judgment about when and how to apply rules, which remains outside reliable AI capability.
Technical feasibility todayclaude-sonnet-51/5No deployed AI product performs institutional rule enforcement or supervisory compliance actions in correctional settings; this remains a human authority function.

Maintain order, discipline, and security within assigned areas in accordance with relevant rules, regulations, policies, and laws.

0

CI 0–0 · exposure 0 · augmentation 25 · importance 4.7/5 · click for rater detail

Panel ratings per dimension for this task
DimensionModelRatingRationale
Sector adoption velocityclaude-haiku-4-5-202510011/5Correctional facilities operate in a heavily regulated, security-sensitive, and labor-intensive sector with slow digitization and deep human-presence requirements. Adoption of AI for supervisory functions is minimal; facilities are still deploying basic monitoring tools, not autonomous management systems.
Sector adoption velocityclaude-sonnet-51/5Corrections is a highly physical, low-digitization public-sector environment with minimal AI agent adoption for direct custodial supervision tasks.
Augmentation potentialclaude-haiku-4-5-202510012/5AI can support supervisors with monitoring dashboards, alert systems, or incident logging, but these tools offer modest productivity gains. The core decision-making and leadership role remains human-driven, limiting the transformative potential of augmentation in this context.
Augmentation potentialclaude-sonnet-52/5AI can assist with monitoring systems, incident logging, or scheduling, but it offers little direct augmentation to the moment-to-moment exercise of authority and physical oversight required.
Task automatabilityclaude-haiku-4-5-202510011/5Maintaining order, discipline, and security in correctional facilities requires real-time judgment, conflict de-escalation, personnel management, and legal accountability that depend on human presence and authority. Current AI systems cannot perform the dynamic decision-making and interpersonal intervention necessary to manage this task at scale.
Task automatabilityclaude-sonnet-51/5This task requires physical presence, real-time judgment during confrontations, and legal authority to enforce discipline within a secure facility; no AI system can perform this end-to-end.
Adoption barriersclaude-haiku-4-5-202510015/5Strong legal and regulatory barriers protect this role: correctional officers and their supervisors must be licensed and authorized by the state; security decisions carry liability exposure; and laws explicitly require human command authority and accountability in correctional settings. No automation can substitute without legal change.
Adoption barriersclaude-sonnet-55/5Correctional supervision requires sworn/certified correctional officers with legal authority to use force, enforce rules, and be accountable under law, making this a hard-barrier, human-only role.
Cost vs. human wageclaude-haiku-4-5-202510011/5This task requires a licensed, salaried first-line supervisor on-site. AI surveillance or decision-support tools add cost without replacing the supervisor; the human cost remains the primary expense and cannot be displaced by current technology.
Cost vs. human wageclaude-sonnet-51/5There is no AI substitute capable of performing this task at all, so cost comparison favors the human by default since AI cannot deliver the output.
Technical feasibility todayclaude-haiku-4-5-202510011/5No deployed AI systems can autonomously maintain correctional facility discipline or security. While sensors and monitoring tools exist, the core task—exercising supervisory judgment and command authority over officers and inmates—remains a human-only function with legal and safety implications that preclude current AI automation.
Technical feasibility todayclaude-sonnet-51/5No deployed AI product supervises inmates, intervenes in altercations, or enforces institutional discipline; this remains entirely a human physical and authority-based function.

Respond to emergencies, such as escapes.

0

CI 0–0 · exposure 0 · augmentation 25 · importance 4.6/5 · click for rater detail

Panel ratings per dimension for this task
DimensionModelRatingRationale
Sector adoption velocityclaude-haiku-4-5-202510011/5Correctional facilities are government-operated, risk-averse, and subject to strict regulatory requirements that mandate human supervisory authority. Adoption of AI for emergency response decision-making is negligible in this sector.
Sector adoption velocityclaude-sonnet-51/5Corrections is a low-digitization, physically grounded sector with minimal AI deployment for operational emergency response; adoption is essentially nonexistent for this function.
Augmentation potentialclaude-haiku-4-5-202510012/5AI could provide limited augmentation through real-time data feeds or alert systems to assist supervisors in situational awareness, but the core emergency response task requires human judgment and authority that AI cannot meaningfully enhance.
Augmentation potentialclaude-sonnet-52/5AI can support situational awareness via surveillance analytics, alarms, or predictive alerts, but it provides only marginal assistance to the actual emergency response and decision-making by supervisors.
Task automatabilityclaude-haiku-4-5-202510011/5Emergency response to escapes requires real-time situational judgment, coordination of personnel, dynamic risk assessment, and physical intervention capabilities that current AI cannot perform autonomously. This task inherently demands human decision-making authority and cannot be delegated to AI systems.
Task automatabilityclaude-sonnet-51/5Responding to a live security emergency like an escape requires physical presence, real-time judgment, and authority to direct personnel and use force decisions; no AI system can execute this end-to-end.dev
Adoption barriersclaude-haiku-4-5-202510015/5Correctional facility operations are heavily regulated and legally mandated to be under human command authority. Legal liability, public safety mandates, and security regulations require licensed, accountable human supervisors to make and sign off on emergency response decisions; this is a hard legal barrier.
Adoption barriersclaude-sonnet-55/5This is a security- and safety-critical function requiring sworn/authorized correctional staff with legal authority to use force and make custodial decisions, making it heavily regulated and non-delegable to AI.
Cost vs. human wageclaude-haiku-4-5-202510011/5The cost of AI systems to monitor and respond to escapes, combined with necessary human oversight and liability exposure, far exceeds the cost of maintaining human supervisory staff whose salaries are already sunk in facility operations.
Cost vs. human wageclaude-sonnet-51/5There is no AI substitute performing this task, so cost comparison favors the human by default; any AI involvement (alerts, monitoring) is only a small input cost, not a substitute for the task itself.
Technical feasibility todayclaude-haiku-4-5-202510011/5No deployed AI product can reliably respond to prison escapes end-to-end or independently. While AI might assist with alert systems or data analysis, the actual emergency response coordination and command authority must remain with human supervisors.
Technical feasibility todayclaude-sonnet-51/5No deployed product performs emergency response and command in a correctional facility; this remains entirely a research-irrelevant, human-only function today.

Supervise and direct the work of correctional officers to ensure the safe custody, discipline, and welfare of inmates.

0

CI 0–0 · exposure 0 · augmentation 25 · importance 4.6/5 · click for rater detail

Panel ratings per dimension for this task
DimensionModelRatingRationale
Sector adoption velocityclaude-haiku-4-5-202510011/5Correctional agencies are among the most conservative and regulated sectors, with minimal technology adoption for core operational tasks. No measurable displacement of supervisory roles by AI is occurring in U.S. prisons or jails.
Sector adoption velocityclaude-sonnet-51/5Corrections is a low-digitization, physically-grounded government sector with minimal AI agent adoption for frontline supervisory authority.
Augmentation potentialclaude-haiku-4-5-202510012/5AI might assist with scheduling, incident logging, or data aggregation, but the core supervisory task of directing officers and making judgment calls in real time is not meaningfully augmented by current systems. Human supervisors would remain fully responsible for all critical decisions.
Augmentation potentialclaude-sonnet-52/5AI can assist with scheduling, incident documentation, or data analysis for staffing patterns, but offers limited help with the core real-time supervisory and safety judgment tasks.
Task automatabilityclaude-haiku-4-5-202510011/5Supervising and directing correctional officers requires real-time human judgment, relationship management, safety assessments, and discretionary decision-making in high-stakes environments. Current AI cannot reliably perform these supervisory duties end-to-end.
Task automatabilityclaude-sonnet-51/5This is an in-person supervisory and safety-critical management role requiring physical presence, real-time judgment, and authority over staff and inmates; no AI system can perform this end-to-end.
Adoption barriersclaude-haiku-4-5-202510015/5Correctional facilities operate under strict regulatory frameworks requiring named human supervisors with legal authority and accountability. State and federal law typically mandate that supervisory positions be held by licensed or certified personnel with direct legal responsibility for custody and safety.
Adoption barriersclaude-sonnet-55/5Correctional supervision involves legal authority, use-of-force oversight, liability, and statutory requirements that a certified human officer must perform and be accountable for.
Cost vs. human wageclaude-haiku-4-5-202510011/5The loaded cost of a first-line supervisor is modest relative to the organization's risk and liability exposure. Any AI system capable of truly supervising officers would require extensive integration, legal setup, and human oversight, making it more expensive than the human supervisor it would replace.
Cost vs. human wageclaude-sonnet-51/5There is no AI substitute performing this supervisory function, so cost comparison favors the human role entirely; AI cannot deliver equivalent output.
Technical feasibility todayclaude-haiku-4-5-202510011/5No deployed AI products perform first-line supervisory functions in correctional settings. This task demands live personnel management, incident response, and accountability that require human authority and legal responsibility.
Technical feasibility todayclaude-sonnet-51/5No deployed product supervises correctional officers or manages inmate safety and discipline; this remains squarely a human management function.

Supervise or perform searches of inmates or their quarters to locate contraband items.

0

CI 0–0 · exposure 0 · augmentation 25 · importance 4.6/5 · click for rater detail

Panel ratings per dimension for this task
DimensionModelRatingRationale
Sector adoption velocityclaude-haiku-4-5-202510011/5Correctional facilities are traditionalist, compliance-heavy organizations with strong personnel requirements. Adoption of AI for core security functions like contraband searches is negligible; facilities remain dependent on human supervisors and officers.
Sector adoption velocityclaude-sonnet-51/5Corrections is a highly physical, low-digitization sector with minimal AI agent deployment for hands-on security tasks; adoption of any automation here is negligible.
Augmentation potentialclaude-haiku-4-5-202510012/5Basic imaging or scanning technology could assist by flagging anomalies for human review, but augmentation is limited because the core task—judgment, physical search, and authority to act—remains human-dependent and heavily regulated.
Augmentation potentialclaude-sonnet-52/5AI-enabled scanners, contraband-detection cameras, or metal/drug detection tools can assist in flagging risks, but the core supervisory search task itself sees limited AI augmentation.
Task automatabilityclaude-haiku-4-5-202510011/5Searching inmates and quarters for contraband requires physical presence, tactile assessment, decision-making about suspicious items in context, and judgment calls that are not automatable with current technology. Current AI systems cannot physically conduct or reliably oversee searches in real correctional environments.
Task automatabilityclaude-sonnet-51/5Physical searches of inmates and cells require in-person presence, physical dexterity, and judgment about concealment that current AI systems cannot perform; this is fundamentally a physical-world task, not information processing.
Adoption barriersclaude-haiku-4-5-202510015/5Corrections is a highly regulated, security-sensitive domain where searches must be conducted by authorized, trained personnel. Legal liability, inmate safety, constitutional protections against unreasonable search, and facility security protocols create hard barriers to non-human execution of this task.
Adoption barriersclaude-sonnet-55/5Correctional facility security, legal authority, and use-of-force/search protocols require sworn, trained correctional officers to conduct and supervise searches, a hard institutional and legal barrier.
Cost vs. human wageclaude-haiku-4-5-202510011/5Deploying any AI-adjacent technology (cameras, imaging systems, alerting) for contraband detection still requires human supervisors and searchers on-site; the marginal cost of automation is high relative to the human labor it might reduce.
Cost vs. human wageclaude-sonnet-51/5There is no AI substitute performing this physical task, so any 'AI cost' comparison is moot; human correctional staff remain the only viable option.
Technical feasibility todayclaude-haiku-4-5-202510011/5No deployed AI product performs inmate or cell searches autonomously or oversees them reliably. This task inherently requires human physical presence and judgment; it exists only in operational/research contexts, not in production AI systems.
Technical feasibility todayclaude-sonnet-51/5No deployed AI product performs physical pat-downs or cell searches; some sensor/scanner tech assists detection but does not replace the supervisory and physical search task itself.

Monitor behavior of subordinates to ensure alert, courteous, and professional behavior toward inmates, parolees, fellow employees, visitors, and the public.

0

CI 0–0 · exposure 0 · augmentation 25 · importance 4.5/5 · click for rater detail

Panel ratings per dimension for this task
DimensionModelRatingRationale
Sector adoption velocityclaude-haiku-4-5-202510011/5Correctional facilities are among the slowest to digitize and adopt autonomous monitoring. Most rely on traditional human supervision, spot checks, and formal reporting—not algorithmic behavioral assessment.
Sector adoption velocityclaude-sonnet-51/5Corrections is a low-digitization, physically embedded government sector with minimal AI adoption for direct supervisory oversight of staff conduct.
Augmentation potentialclaude-haiku-4-5-202510012/5AI could assist marginally by flagging incidents in video footage or summarizing employee reports, but the judgment of whether behavior meets professional standards and the supervisory conversation with staff remain inherently human tasks requiring authority and accountability.
Augmentation potentialclaude-sonnet-52/5AI could assist with logging incidents, flagging complaint patterns, or reviewing camera footage for anomalies, but this only marginally supports the core in-person monitoring and judgment task.
Task automatabilityclaude-haiku-4-5-202510011/5Monitoring subordinate behavior for subjective qualities like 'alert, courteous, and professional' conduct requires real-time observation in complex social environments with nuanced interpersonal judgment. Current AI systems cannot reliably assess these qualitative behavioral standards or intervene with the authority and contextual understanding a supervisor requires.
Task automatabilityclaude-sonnet-51/5This requires in-person observation, judgment about interpersonal conduct, and authority to correct staff behavior in a secure facility—no AI system can perform this end-to-end today.
Adoption barriersclaude-haiku-4-5-202510015/5Hard barriers exist: supervisory authority requires a licensed human supervisor accountable for personnel decisions; labor agreements typically protect workers from algorithmic discipline; liability for false findings is high; and unions strongly resist surveillance-based evaluation of staff conduct.
Adoption barriersclaude-sonnet-55/5Correctional supervision involves legal authority, chain-of-command accountability, and safety/security regulations requiring certified human supervisors on-site.
Cost vs. human wageclaude-haiku-4-5-202510011/5The cost of deploying comprehensive monitoring systems (cameras, AI analysis, integration with HR systems, plus human oversight to validate findings) would exceed the loaded wage of a first-line supervisor performing this oversight function.
Cost vs. human wageclaude-sonnet-51/5There is no viable AI substitute providing this supervisory function, so cost comparison favors the human role entirely.
Technical feasibility todayclaude-haiku-4-5-202510011/5No deployed product today reliably monitors and evaluates staff behavior across diverse correctional settings to ensure professional conduct. Video analysis systems exist but cannot consistently judge subtleties of courtesy, professionalism, or appropriateness in the varied encounters this role demands.
Technical feasibility todayclaude-sonnet-51/5No deployed product monitors correctional staff behavior and enforces professional conduct standards in real facilities; this remains a human supervisory function.

Restrain, secure, or control offenders, using chemical agents, firearms, or other weapons of force as necessary.

0

CI 0–0 · exposure 0 · augmentation 25 · importance 4.5/5 · click for rater detail

Panel ratings per dimension for this task
DimensionModelRatingRationale
Sector adoption velocityclaude-haiku-4-5-202510011/5Correctional facilities are publicly regulated environments with entrenched labor practices and legal constraints. Adoption of AI for force deployment remains negligible, with no measurable displacement in production settings.
Sector adoption velocityclaude-sonnet-51/5Corrections is a low-digitization, physically demanding sector with minimal AI deployment for direct physical control tasks.
Augmentation potentialclaude-haiku-4-5-202510012/5Limited augmentation potential exists: body cameras, threat-detection alerts, or communication aids might slightly assist human decision-making, but the physical control component itself is not meaningfully aided by current AI.
Augmentation potentialclaude-sonnet-52/5AI can assist indirectly via monitoring systems, threat detection, or decision-support alerts, but offers little direct augmentation during the physical act of restraint itself.
Task automatabilityclaude-haiku-4-5-202510011/5This task requires immediate physical presence, real-time threat assessment, and split-second decisions in a dynamic, high-risk environment. No current AI system can operate independently in physical space to restrain individuals or deploy force safely.
Task automatabilityclaude-sonnet-51/5This is a physical, high-stakes use-of-force task requiring real-time human judgment, physical presence, and legal accountability; no AI system can perform or substitute for this today.
Adoption barriersclaude-haiku-4-5-202510015/5Extensive legal and regulatory frameworks mandate that licensed human correctional officers—often with specific certifications—must physically execute force decisions. Liability, criminal accountability, and due process protections create hard legal requirements for human authority and judgment.
Adoption barriersclaude-sonnet-55/5Use of force is tightly regulated, requires certified/sworn personnel, carries major legal liability, and mandates human accountability and authorization, forming a hard barrier to automation.
Cost vs. human wageclaude-haiku-4-5-202510011/5The capital, maintenance, liability insurance, and oversight infrastructure required for autonomous restraint systems far exceeds the fully-loaded cost of trained correctional supervisors.
Cost vs. human wageclaude-sonnet-51/5There is no AI equivalent performing this physical task, so cost comparison is moot; a human officer is the only viable option and thus AI is not cheaper by any measure.
Technical feasibility todayclaude-haiku-4-5-202510011/5No deployed product performs autonomous physical restraint or weapons deployment in correctional settings. Remote teleoperation systems exist in limited research contexts but are not operationally deployed at scale for this critical safety function.
Technical feasibility todayclaude-sonnet-51/5No deployed AI product physically restrains or uses force against people; this remains entirely a human physical and judgment-based action.

Carry injured offenders or employees to safety and provide emergency first aid when necessary.

0

CI 0–0 · exposure 0 · augmentation 13 · importance 4.5/5 · click for rater detail

Panel ratings per dimension for this task
DimensionModelRatingRationale
Sector adoption velocityclaude-haiku-4-5-202510011/5This task occurs in heavily regulated, physically-grounded correctional settings with low digitization and mandatory human supervisory presence; no AI adoption is occurring.
Sector adoption velocityclaude-sonnet-51/5Corrections is a low-digitization, physically demanding sector with minimal robotics or AI adoption for hands-on emergency physical tasks.
Augmentation potentialclaude-haiku-4-5-202510012/5AI could marginally assist by alerting supervisors to health incidents via camera monitoring or biometric systems, but the core task of physically carrying and administering aid remains entirely human-dependent.
Augmentation potentialclaude-sonnet-51/5AI offers essentially no assistance to the physical acts of carrying an injured person or performing first aid in real time.
Task automatabilityclaude-haiku-4-5-202510011/5This task requires physical movement of injured persons, situational assessment in an emergency context, and direct medical intervention—all inherently human activities. Current AI systems cannot physically carry people or provide hands-on first aid.
Task automatabilityclaude-sonnet-51/5This is a physical task requiring bodily strength, mobility, and hands-on emergency response in a secure facility; no AI system can physically carry a person or administer first aid.
Adoption barriersclaude-haiku-4-5-202510015/5Strong legal and regulatory barriers require licensed or trained personnel to provide emergency response and first aid; liability and duty-of-care requirements mandate human presence and responsibility.
Adoption barriersclaude-sonnet-55/5Correctional supervisors are legally required to ensure safety and security, and emergency medical response and physical intervention require certified, authorized human personnel with legal responsibility.
Cost vs. human wageclaude-haiku-4-5-202510011/5AI cannot perform this task, so direct cost comparison is inapplicable; a human supervisor must always be present and capable of physical response.
Cost vs. human wageclaude-sonnet-51/5There is no viable AI substitute, so any hypothetical AI/robotic solution (e.g., rescue robots) would be far more expensive and less capable than a human supervisor for this purpose.
Technical feasibility todayclaude-haiku-4-5-202510011/5No deployed AI system can physically transport injured individuals or administer emergency first aid; this remains entirely human-dependent work in production correctional facilities.
Technical feasibility todayclaude-sonnet-51/5No deployed product performs physical carrying or hands-on first aid; this remains entirely outside current AI/robotics deployment in corrections settings.

Supervise activities, such as searches, shakedowns, riot control, or institutional tours.

0

CI 0–0 · exposure 0 · augmentation 25 · importance 4.3/5 · click for rater detail

Panel ratings per dimension for this task
DimensionModelRatingRationale
Sector adoption velocityclaude-haiku-4-5-202510011/5Correctional facilities are traditionally low-digitization environments with strong regulatory and union constraints. Adoption of AI for supervisory roles in security-critical settings has been minimal, and public sector budgets and liability concerns create substantial lag.
Sector adoption velocityclaude-sonnet-51/5Corrections is a low-digitization, physically-intensive public-sector field with minimal AI agent deployment for direct security supervision tasks.
Augmentation potentialclaude-haiku-4-5-202510012/5AI monitoring systems might assist with detecting anomalies or logging activities in searches or tours, but current systems offer limited meaningful augmentation for the core supervisory judgment and tactical command functions required during dynamic security events.
Augmentation potentialclaude-sonnet-52/5AI could assist with logging, scheduling, or analyzing surveillance footage to support decision-making, but it does not meaningfully enhance the live, physical supervisory judgment required.
Task automatabilityclaude-haiku-4-5-202510011/5Supervision of dynamic, high-stakes correctional security tasks like riot control and shakedowns requires real-time situational awareness, tactical decision-making, and immediate response to unpredictable human behavior that current AI systems cannot reliably perform end-to-end. These activities involve complex judgment calls about safety, use of force, and personnel coordination that demand human accountability and discretion.
Task automatabilityclaude-sonnet-51/5This involves physically supervising high-stakes, safety-critical activities inside a correctional facility, requiring direct human presence, judgment, and authority that no AI system can perform end-to-end.
Adoption barriersclaude-haiku-4-5-202510015/5Correctional facility operations are heavily regulated, and supervision of security activities like riot control involves explicit legal authority and liability that must reside with licensed, accountable human staff. Regulations explicitly require qualified human supervisors to direct and oversee correctional security operations.
Adoption barriersclaude-sonnet-55/5Correctional supervision is governed by strict statutory authority, security protocols, and liability concerns requiring a sworn, trained officer physically present and legally responsible.
Cost vs. human wageclaude-haiku-4-5-202510011/5Supervision of high-risk correctional activities requires human presence for liability, authority, and immediate intervention; AI systems would need to be paired with human supervisors anyway, making the combined cost higher than a human supervisor alone.
Cost vs. human wageclaude-sonnet-51/5There is no AI substitute performing this supervisory function, so cost comparison is moot; any AI cost would be additive to, not a replacement for, the human supervisor.
Technical feasibility todayclaude-haiku-4-5-202510011/5No deployed AI system can autonomously supervise correctional security operations today. While vision systems can detect some anomalies, the integration of real-time tactical oversight, personnel direction, and safety-critical decision-making under duress remains beyond current production capabilities.
Technical feasibility todayclaude-sonnet-51/5No deployed product supervises searches, riot control, or institutional tours; this remains firmly in the domain of trained corrections personnel with no commercial AI analog.

Resolve problems between inmates.

0

CI 0–0 · exposure 0 · augmentation 25 · importance 4.3/5 · click for rater detail

Panel ratings per dimension for this task
DimensionModelRatingRationale
Sector adoption velocityclaude-haiku-4-5-202510011/5Correctional institutions are cautious, heavily regulated sectors with limited AI adoption. The nature of this task—managing human conflict—remains dependent on trained staff presence and authority rather than any technological substitution.
Sector adoption velocityclaude-sonnet-51/5Corrections is a highly regulated, low-digitization, physically-mediated sector with minimal AI adoption for direct inmate conflict resolution.
Augmentation potentialclaude-haiku-4-5-202510012/5AI might marginally assist by documenting incidents or providing background information on conflicts, but the core activity of de-escalation and problem-resolution relies on human presence, authority, and interpersonal skill that AI cannot meaningfully augment.
Augmentation potentialclaude-sonnet-52/5AI could help with monitoring systems (e.g., flagging tension via camera analytics) or paperwork after the fact, but offers little direct assistance during actual conflict resolution.
Task automatabilityclaude-haiku-4-5-202510011/5Resolving problems between inmates requires real-time judgment about interpersonal conflict, safety risk assessment, and understanding complex social dynamics within a correctional setting. Current AI systems lack the situational awareness, physical presence, and ability to make high-stakes decisions that this task demands.
Task automatabilityclaude-sonnet-51/5Resolving interpersonal conflicts between inmates requires physical presence, authority, real-time judgment, and de-escalation skills that AI cannot perform end-to-end today.
Adoption barriersclaude-haiku-4-5-202510015/5Correctional facilities are heavily regulated and require licensed, trained personnel with legal authority to mediate disputes and make decisions about inmate management. Human judgment and presence are legally mandated for this role.
Adoption barriersclaude-sonnet-55/5Legal authority, use-of-force regulations, safety/security mandates, and the requirement for a sworn/trained officer to physically intervene make this a hard, non-negotiable human function.
Cost vs. human wageclaude-haiku-4-5-202510011/5AI solutions cannot perform this task at all, making cost comparison inapplicable; a human supervisor is irreplaceable for this function regardless of financial comparison.
Cost vs. human wageclaude-sonnet-51/5There is no viable AI substitute performing this task, so any AI cost comparison is moot; the human correctional officer remains the only functional option.
Technical feasibility todayclaude-haiku-4-5-202510011/5No deployed AI product exists that can independently resolve conflicts between inmates in a correctional facility. This task fundamentally requires human authority, presence, and de-escalation capability that AI cannot currently provide in production environments.
Technical feasibility todayclaude-sonnet-51/5No deployed AI product mediates or resolves physical altercations or disputes between incarcerated individuals; this remains entirely human-performed.

Transfer or transport offenders on foot or by driving vehicles, such as trailers, vans, or buses.

0

CI 0–0 · exposure 0 · augmentation 25 · importance 4.1/5 · click for rater detail

Panel ratings per dimension for this task
DimensionModelRatingRationale
Sector adoption velocityclaude-haiku-4-5-202510011/5Correctional agencies operate under strict statutory constraints and have minimal incentive or legal permission to automate prisoner transport; adoption velocity is effectively zero.
Sector adoption velocityclaude-sonnet-51/5Corrections is a highly physical, low-digitization sector with essentially no AI adoption for direct custodial/transport functions.
Augmentation potentialclaude-haiku-4-5-202510012/5AI might assist with route optimization or vehicle maintenance tracking, but cannot augment the core security and custody judgment required during offender transport, limiting meaningful productivity gain.
Augmentation potentialclaude-sonnet-52/5AI could assist with route planning, scheduling, or monitoring systems, but offers minimal direct assistance to the physical act of transporting offenders.
Task automatabilityclaude-haiku-4-5-202510011/5Transporting offenders in secure custody requires continuous physical supervision, real-time decision-making in dynamic environments, and immediate response to security incidents. Current AI cannot reliably manage the unpredictable human and safety variables involved in escorting detained individuals.
Task automatabilityclaude-sonnet-51/5Physically escorting and driving offenders requires physical presence, security judgment, and use-of-force capability that no AI system can perform; this is a physical custody task, not an information task.
Adoption barriersclaude-haiku-4-5-202510015/5This task is protected by strong legal and regulatory barriers: only authorized correctional officers can transport offenders in custody; liability for escape or harm is severe; and federal and state law mandate direct human supervision of prisoners in transit.
Adoption barriersclaude-sonnet-55/5Correctional officer duties involving custody and transport of offenders are legally mandated to be performed by authorized, trained, and often armed personnel with statutory authority, creating hard legal and safety barriers.
Cost vs. human wageclaude-haiku-4-5-202510011/5AI cannot substitute for this task; autonomous vehicles cannot legally transport prisoners, and the liability, security, and oversight costs far exceed any theoretical AI cost savings.
Cost vs. human wageclaude-sonnet-51/5There is no AI substitute performing this physical transport/custody function, so AI cost is not comparable—human labor is the only option today.
Technical feasibility todayclaude-haiku-4-5-202510011/5No deployed product autonomously transports offenders. This task fundamentally requires a licensed, trained human operator with legal authority and physical presence for custody and security purposes.
Technical feasibility todayclaude-sonnet-51/5No deployed AI product transports or physically escorts incarcerated individuals; this remains purely a human/physical security function.

Supervise or provide security for offenders performing tasks, such as construction, maintenance, laundry, food service, or other industrial or agricultural operations.

0

CI 0–0 · exposure 0 · augmentation 25 · importance 4.0/5 · click for rater detail

Panel ratings per dimension for this task
DimensionModelRatingRationale
Sector adoption velocityclaude-haiku-4-5-202510011/5Correctional facilities operate under rigid legal and safety mandates with little pressure to automate human-intensive security roles. Adoption of AI tools remains minimal, confined to ancillary tasks like lock control, with human supervision remaining non-negotiable.
Sector adoption velocityclaude-sonnet-51/5Corrections is a slow-moving, highly regulated, physically-grounded government sector with minimal AI adoption for frontline custodial supervision tasks.
Augmentation potentialclaude-haiku-4-5-202510012/5Camera monitoring systems and alert tools can assist supervisors by flagging unusual activity, but the assistant role is narrow and peripheral. The core work—judgment, presence, intervention, and legal accountability—remains fundamentally dependent on the human officer.
Augmentation potentialclaude-sonnet-52/5AI could assist with scheduling, monitoring camera feeds, or flagging anomalies, but it offers little direct augmentation to the core in-person security supervision task itself.
Task automatabilityclaude-haiku-4-5-202510011/5Supervising and providing security for offenders requires continuous real-time monitoring, risk assessment, physical presence, and immediate intervention in unpredictable situations. Current AI systems cannot reliably detect threats, make split-second security decisions, or physically manage incapacitated individuals—all core requirements for this task.
Task automatabilityclaude-sonnet-51/5This requires physical presence, direct observation, and real-time authority over incarcerated individuals performing manual labor; no AI system can perform physical security supervision.
Adoption barriersclaude-haiku-4-5-202510015/5Correctional supervision is strictly regulated; state and federal law typically requires licensed, trained correctional personnel with direct legal responsibility for inmate security. No automation can replace the mandated human authority and legal liability of a correctional supervisor.
Adoption barriersclaude-sonnet-55/5Correctional supervision is heavily regulated, requires sworn/certified officers with legal authority to use force and maintain custody, creating hard legal and institutional barriers to automation.
Cost vs. human wageclaude-haiku-4-5-202510011/5AI systems capable of monitoring operations would require continuous high-quality video analysis, integration with facility management systems, and significant human oversight, making the all-in cost comparable to or exceeding the cost of a human supervisor.
Cost vs. human wageclaude-sonnet-51/5There is no AI substitute performing this physical security role, so cost comparison favors the human by default; AI cannot deliver the output at all.
Technical feasibility todayclaude-haiku-4-5-202510011/5No deployed AI system can substitute for human security supervision of incarcerated populations. While computer vision can assist with monitoring, actual custody, threat detection, and security decision-making remain legally and operationally dependent on human officers.
Technical feasibility todayclaude-sonnet-51/5No deployed AI product provides physical custodial supervision or security oversight of offenders in work settings; this remains a human-only function in practice.

Related occupations — Protective Service

How to read this

A high substitution score does not mean this job disappears — it means a large share of its current tasks face replacement pressure, so the mix of tasks is likely to change. High augmentation alongside substitution typically means the occupation reorganizes around the protected tasks. Wide confidence intervals mean the rater panel disagreed: treat those scores as open questions, not verdicts.

What would change this score

New model capabilities (automatability, feasibility), falling inference costs (cost ratio), regulation and licensing shifts (barriers), and measured sector adoption (velocity) all re-enter at every index release. Each release is recomputed, versioned and kept queryable — scores are claims with a date on them, not permanent labels.