Correctional Officers and Jailers

33-3012.00
Median wage $58,940/yr380,500 employed (US)Rank #864 of 923 scored · top 94% by substitution

Guard inmates in penal or rehabilitative institutions in accordance with established regulations and procedures. May guard prisoners in transit between jail, courtroom, prison, or other point. Includes deputy sheriffs and police who spend the majority of their time guarding prisoners in correctional institutions.

Sub-scores

0–100 · band = confidence interval from rater disagreement

Substitution11
Exposure12
Augmentation36

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

27 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%11

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

Technical feasibility todayw 20%13

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

Cost vs. human wagew 15%14

panel mean rating 1.5/5 → substitution pressure 14/100

Adoption barriersw 20%inverted — strong barriers lower the score9

panel mean rating 4.6/5 (barrier strength) → substitution pressure 9/100

Sector adoption velocityw 10%6

panel mean rating 1.2/5 → substitution pressure 6/100

Task breakdown (27 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.

Maintain records of prisoners' identification and charges.

62

CI 56–69 · exposure 70 · augmentation 75 · importance 4.4/5 · click for rater detail

Panel ratings per dimension for this task
DimensionModelRatingRationale
Sector adoption velocityclaude-haiku-4-5-202510012/5Correctional facilities lag in digital transformation compared to finance or tech; many still rely on paper-based or legacy systems. Adoption of AI-driven record management is emerging but slow, constrained by budget, legacy infrastructure, and organizational change resistance.
Sector adoption velocityclaude-sonnet-52/5Government and corrections sectors are historically slow adopters of new technology due to budget constraints, legacy systems, and procurement processes, resulting in limited but growing digitization.
Augmentation potentialclaude-haiku-4-5-202510014/5AI can substantially assist officers by auto-populating records, cross-checking data for consistency, and flagging missing information, allowing human staff to focus on verification and legal accuracy rather than manual transcription.
Augmentation potentialclaude-sonnet-54/5AI-assisted data entry, OCR for ID scanning, and automated record cross-referencing can meaningfully speed up and reduce errors in this task while officers remain responsible for final verification.
Task automatabilityclaude-haiku-4-5-202510014/5Current AI systems can reliably extract, organize, and maintain structured records of prisoner identification and charges from documents, forms, and databases with minimal human intervention. This is primarily data entry, classification, and record-keeping—tasks where AI achieves well over 50% time savings at equal quality.
Task automatabilityclaude-sonnet-54/5Records maintenance is largely structured data entry and retrieval, which off-the-shelf database and OCR/document systems can handle with significant time savings, though some manual verification remains.
Adoption barriersclaude-haiku-4-5-202510013/5Some regulatory requirements around data integrity and chain-of-custody documentation, plus institutional inertia in correctional systems, create moderate friction. However, no strict legal requirement mandates a human perform this task, only that records be accurate and maintained.
Adoption barriersclaude-sonnet-53/5No licensing requirement mandates a specific human to maintain records, but chain-of-custody, data accuracy for legal proceedings, and government procurement/security requirements create moderate friction.
Cost vs. human wageclaude-haiku-4-5-202510015/5AI-powered record management is orders of magnitude cheaper than human data entry and transcription once set up, with minimal per-record inference cost and no ongoing wage expense for routine updates and lookups.
Cost vs. human wageclaude-sonnet-53/5Software licensing and integration costs for records systems are moderate; while cheaper than dedicated clerical staff at scale, correctional facilities still need trained personnel for data accuracy and legal compliance, keeping costs roughly comparable.
Technical feasibility todayclaude-haiku-4-5-202510014/5Deployed products (document processing, OCR, database management systems with AI assistance) reliably perform prisoner record digitization and maintenance in correctional facilities today. Error rates are low for structured data (names, charges, IDs), though complex legal descriptions may require human review.
Technical feasibility todayclaude-sonnet-53/5Jail management systems already digitize booking and charge records, but full automation of intake data entry, verification against ID documents, and integration across jurisdictions still requires human oversight in most facilities.

Inspect mail for the presence of contraband.

34

CI 25–43 · exposure 38 · augmentation 63 · importance 4.4/5 · click for rater detail

Panel ratings per dimension for this task
DimensionModelRatingRationale
Sector adoption velocityclaude-haiku-4-5-202510012/5Correctional facilities are traditionally low-digitization, risk-averse sectors with slow procurement cycles and limited budgets. While some larger facilities use X-ray screening, widespread adoption of AI-driven contraband detection remains in early pilot stages, not production deployment at scale.
Sector adoption velocityclaude-sonnet-52/5Corrections is a slow-moving, budget-constrained public sector with limited digitization and cautious technology adoption, resulting in gradual and inconsistent uptake of scanning technologies.
Augmentation potentialclaude-haiku-4-5-202510013/5AI-assisted image flagging (highlighting suspicious items in scans or X-rays) can moderately improve officer efficiency and consistency in reviewing mail. However, the human officer remains the final decision-maker, and augmentation is limited to narrowing the search space rather than transforming productivity.
Augmentation potentialclaude-sonnet-54/5AI-enhanced imaging and detection tools meaningfully speed up and improve accuracy of manual mail screening, allowing officers to focus attention on flagged suspicious items.
Task automatabilityclaude-haiku-4-5-202510012/5AI could assist with identifying obvious contraband in scanned mail images (drugs, weapons, metal), but the task requires judgment about intent, context, and legal boundaries that vary by facility and jurisdiction. Current systems have high error rates on ambiguous items and cannot reliably detect all contraband types, falling well short of the 50% time-saving threshold.
Task automatabilityclaude-sonnet-53/5AI-based scanning (X-ray, chemical detection, imaging for narcotics-laced paper) can automate detection of many contraband types, but final judgment and physical handling still require human involvement, so only partial time savings are achievable end-to-end.
Adoption barriersclaude-haiku-4-5-202510014/5Correctional facilities operate under strict liability and safety regulations; missing contraband can result in security breaches, escapes, or deaths. These high error costs, combined with facility protocols, chain-of-custody requirements, and the legal accountability placed on human officers, create strong institutional and regulatory friction against full automation.
Adoption barriersclaude-sonnet-54/5Legal and security requirements mandate documented chain-of-custody, human verification, and accountability for contraband findings, creating strong institutional and liability-driven barriers to full automation.
Cost vs. human wageclaude-haiku-4-5-202510012/5High-quality X-ray and imaging systems, plus AI integration and human oversight, are expensive to deploy and maintain. The loaded cost of a correctional officer's mail inspection time is relatively low, making the all-in cost of AI automation comparable or higher when accounting for infrastructure and false-positive review overhead.
Cost vs. human wageclaude-sonnet-53/5Automated scanning equipment and digital mail conversion services have real capital and subscription costs comparable to labor savings, not a clear order-of-magnitude reduction once oversight and equipment maintenance are factored in.
Technical feasibility todayclaude-haiku-4-5-202510012/5Computer vision systems exist for mail screening in some facilities, but they are typically narrow (metal detection, X-ray analysis) rather than end-to-end contraband identification. No mature, general-purpose AI product reliably replaces human mail inspection in production at scale across the range of contraband types and contextual judgments required.
Technical feasibility todayclaude-sonnet-53/5Some correctional facilities deploy body scanners, mail-scanning software, and drug-detection systems from vendors like RAPISCAN or Intelligent Contraband Control, but coverage is uneven and many facilities still rely on manual inspection or outsourced digital mail scanning.

Use nondisciplinary tools and equipment, such as a computer.

33

CI 0–66 · exposure 38 · augmentation 38 · importance 4.3/5 · click for rater detail

Panel ratings per dimension for this task
DimensionModelRatingRationale
Sector adoption velocityclaude-haiku-4-5-202510011/5This task statement describes generic tool use, not a process that sectors adopt or avoid. Correctional facilities are low-digitization, security-sensitive environments where human officers remain mandatory.
Sector adoption velocityclaude-sonnet-52/5Corrections is a slow-adopting, government-run, low-digitization sector where legacy systems and budget constraints limit deep AI integration despite basic computer use being standard.
Augmentation potentialclaude-haiku-4-5-202510012/5While AI might assist with specific software features (e.g., predictive alerts in a case-management system), the vague statement about using computers generically offers minimal basis for meaningful augmentation scoring.
Augmentation potentialclaude-sonnet-53/5AI-enabled software (auto-fill forms, search, alerts) can meaningfully speed up routine computer tasks like report writing and record lookup for officers.
Task automatabilityclaude-haiku-4-5-202510011/5Using nondisciplinary tools and equipment like computers is a generic capability that does not constitute a discrete occupational task requiring automation. The statement lacks specific task objectives (e.g., scheduling, record-keeping) that would enable end-to-end AI automation.
Task automatabilityclaude-sonnet-54/5Basic computer use for logging, reporting, and record-keeping is highly automatable with existing software, forms automation, and AI-assisted data entry tools.reaching the 50% time-saving bar for routine administrative computer tasks.
Adoption barriersclaude-haiku-4-5-202510015/5Correctional officers must be present and licensed/hired by the institution; their role requires direct human oversight, physical presence, and legal accountability that cannot be delegated to machines.
Adoption barriersclaude-sonnet-52/5No licensing requirement dictates that only a human perform basic computer tasks, though facility security protocols and access-control policies create some institutional friction.
Cost vs. human wageclaude-haiku-4-5-202510011/5Cost comparison is not meaningful when the task is not automatable. The human must operate the computer as part of their broader correctional duties.
Cost vs. human wageclaude-sonnet-53/5Software licensing and IT support costs are moderate; while cheaper than manual paperwork over time, the officer must still operate the system, so savings are incremental rather than order-of-magnitude.
Technical feasibility todayclaude-haiku-4-5-202510011/5This is not a task that AI systems perform independently; it describes basic tool operation that humans use as inputs to their actual work. No product autonomously 'uses a computer' as a standalone occupational task.
Technical feasibility todayclaude-sonnet-54/5Correctional facilities widely deploy jail management systems, digital logging software, and administrative tools already; these are mature, deployed products, though not AI-specific.

Record information, such as prisoner identification, charges, and incidents of inmate disturbance, keeping daily logs of prisoner activities.

30

CI 23–37 · exposure 33 · augmentation 63 · importance 4.5/5 · click for rater detail

Panel ratings per dimension for this task
DimensionModelRatingRationale
Sector adoption velocityclaude-haiku-4-5-202510011/5Correctional agencies are traditionally low-digitization, risk-averse organizations with slow technology adoption and aging IT infrastructure. Deployment of AI for official record-keeping in this sector remains minimal, with most facilities still relying on manual or basic digital logs.
Sector adoption velocityclaude-sonnet-52/5Corrections is a slow-adopting, highly regulated public-sector environment with limited digitization and cautious rollout of new tech due to security and liability concerns.
Augmentation potentialclaude-haiku-4-5-202510013/5AI could assist by auto-populating structured fields (identification numbers, charges from intake forms) and flagging potential incident keywords for officer review, materially speeding data entry while officers retain judgment on classification and incident assessment.
Augmentation potentialclaude-sonnet-54/5AI-assisted transcription, templated report generation, and automated flagging of keywords (e.g., 'disturbance') can meaningfully speed up documentation while officers verify and finalize records.
Task automatabilityclaude-haiku-4-5-202510012/5While data entry components could be partially automated (e.g., parsing names and charges from documents), the task requires judgment in determining what constitutes a reportable 'incident of inmate disturbance' and contextual interpretation that would require significant human oversight. Current AI cannot reliably end-to-end automate this without substantial human re-verification.
Task automatabilityclaude-sonnet-53/5Transcription, categorization, and log entry could largely be automated via speech-to-text and structured data entry tools, but incident interpretation and judgment on disturbances still require human input for accuracy and accountability.
Adoption barriersclaude-haiku-4-5-202510014/5Correctional facilities operate under regulatory frameworks requiring documented accountability and often have explicit custody-documentation standards. Prison liability for incidents creates strong asymmetric error costs, and legal/audit requirements typically demand human sign-off on official records, creating hard barriers to full automation.
Adoption barriersclaude-sonnet-54/5Recordkeeping in custody settings is subject to strict legal, chain-of-custody, and security clearance requirements, often mandating certified correctional staff to verify and sign off on entries.
Cost vs. human wageclaude-haiku-4-5-202510012/5The cost of AI systems (forms, integration, error review, liability oversight) combined with mandatory human verification of incident classification approaches or exceeds the loaded wage of entry-level correctional staff performing this routine documentation task.
Cost vs. human wageclaude-sonnet-53/5Basic transcription/logging software is cheap, but integration with secure jail management systems and required human verification keeps overall cost comparable to current administrative labor.
Technical feasibility todayclaude-haiku-4-5-202510012/5No mature production system reliably records and categorizes correctional incidents autonomously. While OCR and basic data entry tools exist, they struggle with handwritten logs and nuanced incident classification in real correctional environments, making reliable end-to-end deployment limited to highly structured settings.
Technical feasibility todayclaude-sonnet-52/5Some jail management systems use digital forms and voice-to-text, but reliable automated capture of nuanced incident narratives is not widely deployed in production corrections settings.

Arrange daily schedules for prisoners, including library visits, work assignments, family visits, and counseling appointments.

30

CI 23–37 · exposure 33 · augmentation 50 · importance 3.5/5 · click for rater detail

Panel ratings per dimension for this task
DimensionModelRatingRationale
Sector adoption velocityclaude-haiku-4-5-202510012/5Prisons are traditionally low-digitization, risk-averse organizations with legacy systems; while some facilities use basic scheduling software, autonomous AI scheduling is not in widespread adoption, and organizational and budgetary constraints slow pilot and rollout pace.
Sector adoption velocityclaude-sonnet-51/5Corrections is a highly regulated, low-digitization government sector with historically slow technology adoption, especially for tasks involving security-sensitive inmate management.
Augmentation potentialclaude-haiku-4-5-202510013/5AI-assisted scheduling tools can help correctional staff generate candidate schedules, flag conflicts, and highlight resource bottlenecks, improving efficiency and reducing manual planning burden, but staff remain firmly in the loop for final approval and adjustment.
Augmentation potentialclaude-sonnet-53/5AI-based scheduling tools can help officers organize and optimize daily schedules by flagging conflicts and suggesting slots, offering moderate productivity gains while officers retain final decision-making authority.
Task automatabilityclaude-haiku-4-5-202510012/5While scheduling software can automate parts of the logistics (time-slot assignment, conflict detection), the task requires contextual knowledge of individual prisoner profiles, behavioral risks, security protocols, and dynamic operational constraints that current AI systems cannot reliably integrate end-to-end without substantial human oversight and adjustment.
Task automatabilityclaude-sonnet-53/5Scheduling logic itself is well-suited to software, but integrating constraints from security classifications, staffing, legal visitation rights, and real-time facility changes requires human judgment and oversight, limiting full end-to-end automation today.
Adoption barriersclaude-haiku-4-5-202510014/5Correctional staff are bound by security protocols, legal liability for inmate safety and due-process rights, and organizational requirements for human accountability; a fully autonomous scheduling system would face regulatory resistance and liability concerns if a security breach or rights violation arose.
Adoption barriersclaude-sonnet-54/5Correctional facilities operate under strict security, legal, and regulatory requirements; decisions affecting prisoner movement, visitation, and privileges typically require staff authorization and accountability, creating strong institutional and liability barriers to full automation.
Cost vs. human wageclaude-haiku-4-5-202510012/5Existing scheduling software requires significant setup, maintenance, and human staff time to oversee outputs; the cost of an integrated AI scheduling system likely exceeds the labor savings from reducing one correctional officer's scheduling work, especially in smaller facilities.
Cost vs. human wageclaude-sonnet-53/5A scheduling algorithm or AI tool could be cheap to run, but the human oversight, security verification, and system customization needed for correctional settings likely bring costs closer to parity with existing officer/administrative time.
Technical feasibility todayclaude-haiku-4-5-202510012/5Scheduling tools exist in prisons but primarily handle mechanical slot-filling; they do not autonomously handle the full complexity of prisoner-specific restrictions, security separations, counselor availability, and real-time schedule changes without human correctional staff verification and manual adjustment.
Technical feasibility todayclaude-sonnet-52/5Generic scheduling/calendaring software exists broadly, but purpose-built AI systems reliably managing prisoner scheduling with security and legal constraints in production correctional facilities are not well documented or widely deployed.

Provide to supervisors oral and written reports of the quality and quantity of work performed by inmates, inmate disturbances and rule violations, and unusual occurrences.

20

CI 18–23 · exposure 20 · augmentation 50 · importance 4.1/5 · click for rater detail

Panel ratings per dimension for this task
DimensionModelRatingRationale
Sector adoption velocityclaude-haiku-4-5-202510011/5Correctional systems are typically laggards in AI adoption, with many facilities operating legacy systems and limited digitization. Budget constraints, resistance to technology in traditionally conservative institutional settings, and the high stakes of error adoption velocity remains very low.
Sector adoption velocityclaude-sonnet-51/5Corrections is a slow-adopting, highly regulated, low-digitization sector with minimal AI deployment in daily custodial reporting workflows.
Augmentation potentialclaude-haiku-4-5-202510013/5AI could usefully assist by auto-formatting incident summaries, flagging keywords from officer notes, or organizing data chronologically, thereby reducing clerical burden. However, the human officer must retain responsibility for content, interpretation, and accuracy given the legal and safety stakes involved.
Augmentation potentialclaude-sonnet-53/5AI writing tools can help officers structure and clean up written reports from notes, offering moderate assistance, though the core observation and oral reporting remain manual.
Task automatabilityclaude-haiku-4-5-202510012/5While AI could assist in summarizing routine incident data and formatting reports, this task fundamentally requires human judgment to assess work quality, interpret inmate behavior, and determine severity of violations. Current systems cannot reliably replace the observational and contextual understanding that a correctional officer brings to determining what constitutes a significant disturbance or unusual occurrence.
Task automatabilityclaude-sonnet-52/5Drafting written reports could be assisted by AI, but the underlying observation, judgment about rule violations, and oral reporting to supervisors require direct human presence and cannot be automated end-to-end.
Adoption barriersclaude-haiku-4-5-202510014/5Significant barriers exist: correctional officers must document incidents accurately for legal liability and security purposes, supervisors rely on human credibility and sign-off for accountability, and any error in incident reporting could create safety or legal exposure. Many facilities also operate under strict authorization protocols for record-keeping and chain-of-custody requirements.
Adoption barriersclaude-sonnet-54/5Security, legal, and institutional accountability requirements mean officers themselves must observe and attest to incidents; reports often have legal/disciplinary consequences requiring a credentialed human author.
Cost vs. human wageclaude-haiku-4-5-202510012/5Implementation would require custom integration with facility management systems, staff training, and ongoing oversight to verify accuracy. The all-in cost of such a system would likely exceed or match the cost of a correctional officer writing reports, particularly given the small scale of many facilities and liability concerns.
Cost vs. human wageclaude-sonnet-52/5AI could cheaply help draft report text once facts are provided, but the costly parts—monitoring, judgment, and oral communication—still require the officer, limiting overall savings.
Technical feasibility todayclaude-haiku-4-5-202510012/5No deployed products currently perform this task end-to-end in correctional facilities. While general summarization and report-generation tools exist, they lack integration with correctional workflows, incident classification systems, and the ability to capture nuanced behavioral observations that supervisors require for safety and operational decisions.
Technical feasibility todayclaude-sonnet-51/5No deployed correctional-facility product autonomously observes inmate behavior and generates authoritative incident reports; this remains a human observational and communication task.

Participate in required job training.

14

CI 0–29 · exposure 13 · augmentation 38 · importance 4.1/5 · click for rater detail

Panel ratings per dimension for this task
DimensionModelRatingRationale
Sector adoption velocityclaude-haiku-4-5-202510011/5Correctional training remains a laggard sector with strong institutional and legal requirements for human instructor presence; no production adoption of AI-led mandatory training exists in this domain.
Sector adoption velocityclaude-sonnet-52/5Corrections is a public-sector, low-digitization field with slow technology adoption; while some agencies use e-learning, comprehensive AI-driven training adoption is limited and pilots are uncommon.
Augmentation potentialclaude-haiku-4-5-202510012/5AI could provide supplementary study materials, scenario simulations, or knowledge-check tools, but the core participatory training requirement and instructor evaluation cannot be meaningfully augmented by current AI systems.
Augmentation potentialclaude-sonnet-53/5AI can supplement training via personalized study modules, scenario simulations, and knowledge testing, aiding retention and scheduling even though it doesn't replace mandatory in-person components.
Task automatabilityclaude-haiku-4-5-202510011/5Job training is inherently human-centered and often involves hands-on demonstrations, role-playing, legal liability instruction, and interpersonal feedback. Current AI cannot meaningfully replace the participatory and evaluative components required for correctional officer certification and compliance training.
Task automatabilityclaude-sonnet-52/5AI can deliver e-learning content and quizzes but the physical, procedural, and situational training (use of force, restraint techniques, facility-specific protocols) requires hands-on human instruction and evaluation.
Adoption barriersclaude-haiku-4-5-202510015/5Correctional officer training is heavily regulated and requires state certification and licensed instructor sign-off; most jurisdictions legally mandate in-person participation and evaluation by authorized trainers, creating a hard regulatory barrier to AI substitution.
Adoption barriersclaude-sonnet-54/5Correctional officer training is often mandated by state law/accreditation bodies requiring certified instructors and in-person qualification (e.g., firearms, use-of-force), creating strong regulatory barriers to full automation.
Cost vs. human wageclaude-haiku-4-5-202510011/5Training participation requires human instructor oversight, facilities, and certification sign-off; AI cannot reduce these costs meaningfully since the human trainer and facility requirements remain mandatory.
Cost vs. human wageclaude-sonnet-53/5AI-delivered training modules can reduce costs for classroom instruction portions, but certified trainers and in-person practical assessments still require paid staff, keeping overall costs comparable rather than dramatically lower.
Technical feasibility todayclaude-haiku-4-5-202510011/5While AI can deliver informational content, no deployed product reliably replaces the participation requirement itself—the trainee must actively engage in exercises, demonstrations, and assessments that are legally and professionally mandated to be supervised by authorized instructors.
Technical feasibility todayclaude-sonnet-52/5E-learning platforms and some AI-driven training content exist and are used in corrections training programs, but they cover only the didactic portion, not the practical/physical skill certification required.

Take fingerprints of arrestees, prisoners, or the general public.

14

CI 9–19 · exposure 9 · augmentation 50 · importance 4.1/5 · click for rater detail

Panel ratings per dimension for this task
DimensionModelRatingRationale
Sector adoption velocityclaude-haiku-4-5-202510013/5Many correctional facilities have upgraded to digital fingerprint scanners and integrated booking systems, but adoption is uneven across jurisdictions and many still use traditional methods. Digitization is moderate and piecemeal rather than rapid.
Sector adoption velocityclaude-sonnet-51/5Corrections is a physically-oriented, security-sensitive, low-digitization sector with slow uptake of AI-driven process automation for custodial tasks.
Augmentation potentialclaude-haiku-4-5-202510013/5Digital fingerprint capture systems assist officers by automating pattern storage, matching against databases, and reducing manual transcription—improving speed and accuracy—while the officer remains responsible for physical capture and verification.
Augmentation potentialclaude-sonnet-53/5Digital fingerprint capture and AI-based matching/identification systems already assist officers by speeding up record lookup and verification, even though the physical fingerprinting act remains manual.
Task automatabilityclaude-haiku-4-5-202510011/5Fingerprint capture requires physical contact with an individual's hands and placement on a scanner or ink surface—tasks that current AI systems cannot perform without robotics, which are not deployed in carceral settings today. The task is fundamentally embodied and cannot be automated end-to-end by software alone.
Task automatabilityclaude-sonnet-51/5Fingerprinting requires physical manipulation of a person's hands against a scanner or ink pad, a hands-on physical task that current AI systems cannot perform without robotic embodiment, which is not deployed for this purpose.
Adoption barriersclaude-haiku-4-5-202510015/5Strong legal and regulatory barriers exist: fingerprinting of arrestees is a procedurally mandated step in criminal processing that typically requires documented chain-of-custody oversight by a sworn officer. Liability, consent, and accuracy verification all require human accountability.
Adoption barriersclaude-sonnet-54/5Custody, chain-of-evidence, and security protocols require sworn correctional/law enforcement personnel to handle intake procedures like fingerprinting, and legal/procedural requirements around identification of arrestees create strong institutional barriers to automation.
Cost vs. human wageclaude-haiku-4-5-202510012/5Digital fingerprint capture equipment and maintenance costs are non-trivial, but the primary expense remains the human correctional officer's labor time, which remains necessary regardless of technology. Cost savings are marginal since human supervision is mandatory.
Cost vs. human wageclaude-sonnet-52/5While digital fingerprint scanners reduce processing time versus ink methods, a trained officer must still be present to conduct the procedure, so there is no full labor substitution and thus limited cost savings from AI itself.
Technical feasibility todayclaude-haiku-4-5-202510012/5While fingerprint scanners and biometric systems exist in production in some facilities, these are hardware devices that still require a human operator to position the arrestee's fingers and manage the capture process. No deployed system performs the entire task autonomously.
Technical feasibility todayclaude-sonnet-52/5Automated fingerprint scanning hardware and matching software (AFIS) exist and are widely used, but the physical act of positioning and capturing prints from a potentially non-cooperative subject still requires a human officer to operate and supervise the process.

Monitor conduct of prisoners in housing unit, or during work or recreational activities, according to established policies, regulations, and procedures, to prevent escape or violence.

12

CI 7–16 · exposure 9 · augmentation 50 · importance 4.6/5 · click for rater detail

Panel ratings per dimension for this task
DimensionModelRatingRationale
Sector adoption velocityclaude-haiku-4-5-202510012/5Correctional facilities are risk-averse, unionized, and slow to adopt technology that reduces staffing. Pilot programs exist, but production deployment of autonomous conduct monitoring is minimal; most facilities still rely on direct human observation as the primary control.
Sector adoption velocityclaude-sonnet-52/5Corrections is a slow-moving, publicly funded, physically-oriented sector with limited AI adoption beyond pilot surveillance-analytics tools; deep production-level agentic adoption is rare.
Augmentation potentialclaude-haiku-4-5-202510013/5AI-assisted video analytics can help officers flag zones of concern or review recorded incidents more efficiently, raising their ability to document and analyze behavior patterns. However, the core task of real-time presence and discretionary intervention remains human-dependent.
Augmentation potentialclaude-sonnet-53/5AI-enhanced camera analytics, motion/altercation detection, and alert systems can help officers notice incidents faster or monitor more areas simultaneously, improving situational awareness without replacing the officer's role.
Task automatabilityclaude-haiku-4-5-202510011/5This task requires real-time physical presence, judgment of subtle behavioral cues, and immediate intervention capacity that current AI systems cannot provide. Video monitoring alone cannot detect the nuanced precursors to violence or escape attempts with sufficient reliability to meet a 50% time-saving threshold for human oversight.
Task automatabilityclaude-sonnet-51/5This requires physical presence, real-time physical intervention capability, and split-second judgment about human behavior and safety in a physical space, which no AI system can perform end-to-end today.'
Adoption barriersclaude-haiku-4-5-202510014/5Correctional officers bear legal and personal responsibility for prisoner safety and security; liability for harm from automation failure is asymmetric and severe. Many jurisdictions have regulatory or contractual requirements for human officer presence in housing units, and the duty of care is difficult to delegate to automated systems.
Adoption barriersclaude-sonnet-55/5Correctional officer duties are governed by strict legal, regulatory, and staffing requirements mandating certified human officers physically present for custody and safety, with significant liability for failures.
Cost vs. human wageclaude-haiku-4-5-202510012/5Surveillance infrastructure (cameras, sensors, integration) requires significant capital and ongoing maintenance. When combined with necessary human review and intervention, the total cost per unit of monitoring remains comparable to or exceeds direct human officer presence.
Cost vs. human wageclaude-sonnet-51/5AI camera systems add cost on top of, not instead of, human officers who must still be physically present to respond, intervene, and maintain custody, so there is no all-in cost savings versus the human role.
Technical feasibility todayclaude-haiku-4-5-202510012/5While video analytics systems exist to flag unusual activity in correctional facilities, they suffer from high false-positive rates and cannot reliably predict violence or escape intent. No deployed product reliably performs the full scope of conduct monitoring and intervention without substantial human oversight.
Technical feasibility todayclaude-sonnet-52/5Video surveillance and some AI-based anomaly/violence detection tools exist in prisons, but they only flag incidents for human response rather than autonomously monitoring or intervening, so deployed capability is narrow and supplementary.

Conduct fire, safety, and sanitation inspections.

11

CI 0–23 · exposure 13 · augmentation 38 · importance 4.2/5 · click for rater detail

Panel ratings per dimension for this task
DimensionModelRatingRationale
Sector adoption velocityclaude-haiku-4-5-202510011/5Correctional facilities are lower-digitization environments with strong organizational and regulatory preference for human accountability. AI adoption in prison operations remains minimal; most facilities conduct inspections via standard human protocols without automation pilots.
Sector adoption velocityclaude-sonnet-51/5Corrections is a highly physical, low-digitization sector with minimal AI agent deployment for facility inspection tasks, showing negligible adoption momentum.
Augmentation potentialclaude-haiku-4-5-202510013/5AI-assisted visual anomaly detection (e.g., flagging potential hazards on video feeds or photos) could help officers identify concerns more quickly, but augmentation is limited by the need for human judgment in code compliance and documented sign-off.
Augmentation potentialclaude-sonnet-52/5AI could help generate checklists, schedule inspections, or flag anomalies from sensor data, but it offers limited direct assistance to the hands-on inspection process itself.
Task automatabilityclaude-haiku-4-5-202510012/5Fire, safety, and sanitation inspections require identifying visual anomalies, hazards, and code violations in complex spatial environments. While AI vision systems could flag some obvious defects, inspections demand contextual judgment about regulatory compliance, severity assessment, and prioritization—tasks that remain largely manual and human-dependent today.
Task automatabilityclaude-sonnet-51/5This requires physical presence, walking through cell blocks, visually inspecting conditions, and exercising judgment about security risks that current AI cannot perform end-to-end.'
Adoption barriersclaude-haiku-4-5-202510015/5Fire, safety, and sanitation compliance in correctional facilities is heavily regulated; inspections typically must be documented by licensed/authorized personnel (correctional staff or certified inspectors) and findings are often legally admissible, creating hard liability and regulatory barriers to full automation.
Adoption barriersclaude-sonnet-54/5Correctional facilities have strict security, custody, and regulatory requirements requiring trained, authorized officers to conduct inspections, creating strong institutional and legal barriers to automation.
Cost vs. human wageclaude-haiku-4-5-202510012/5AI vision hardware, integration, and oversight still cost significantly per inspection cycle; human correctional officers conducting routine inspections as part of their duties carry lower marginal cost than purpose-built AI systems with required validation by qualified personnel.
Cost vs. human wageclaude-sonnet-51/5There is no viable AI substitute performing the physical inspection, so the cost comparison favors the human by default; AI cannot yet replace the labor at any lower cost.
Technical feasibility todayclaude-haiku-4-5-202510012/5Computer vision systems exist for detecting some hazards (e.g., obstructions, structural damage), but deployed products for comprehensive facility inspection remain narrow and require substantial human oversight to validate findings and ensure compliance accuracy. No production systems reliably replace human inspectors end-to-end.
Technical feasibility todayclaude-sonnet-51/5No deployed product autonomously conducts physical fire, safety, and sanitation inspections inside correctional facilities; this remains a human, boots-on-the-ground task.

Conduct head counts to ensure that each prisoner is present.

9

CI 0–18 · exposure 13 · 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 facilities are heavily regulated, slow-moving institutions with minimal AI adoption for core security functions. Head counts remain a human-mandated, non-negotiable security procedure with no sector-wide AI displacement.
Sector adoption velocityclaude-sonnet-51/5Corrections is a slow-adopting, highly regulated, physically-oriented sector with limited AI deployment for core custody functions like headcounts.
Augmentation potentialclaude-haiku-4-5-202510012/5AI-assisted tools like facial recognition or automated record-matching could modestly speed verification against rosters, but the task fundamentally requires human presence and accountability, limiting meaningful augmentation.
Augmentation potentialclaude-sonnet-53/5Electronic tracking systems (RFID, biometric scanners, camera analytics) can assist officers by flagging discrepancies or automating logging, improving speed and accuracy of headcount verification.
Task automatabilityclaude-haiku-4-5-202510011/5Head counts require physical presence verification of specific individuals in a secured facility, demanding real-time human verification and judgment about identity and legitimacy. Current AI cannot reliably identify prisoners across dynamic facility conditions, verify actual presence vs. records, or handle the security implications of errors.
Task automatabilityclaude-sonnet-52/5Physical head counts require presence in a secure facility to visually confirm each inmate; while sensor/RFID/biometric systems exist, they are not yet standard replacements for officer-verified counts in most facilities.dependent on physical verification.pathworks
Adoption barriersclaude-haiku-4-5-202510015/5Correctional facilities operate under strict legal and regulatory mandates requiring a uniformed officer to physically verify prisoner presence and maintain chain-of-custody accountability. Security, liability, and constitutional protections all legally require human certification of prisoner counts.
Adoption barriersclaude-sonnet-55/5Security, legal liability, and regulatory standards in corrections require certified human officers to conduct and verify headcounts; failure has severe safety and legal consequences, making this a hard barrier task.
Cost vs. human wageclaude-haiku-4-5-202510011/5The loaded cost of a correctional officer conducting head counts (part of their hourly wage and security responsibility) is lower than deploying, maintaining, and overseeing automated biometric or surveillance systems with required redundancy and legal compliance.
Cost vs. human wageclaude-sonnet-52/5Sensor/RFID infrastructure requires significant capital investment, installation, and maintenance, and human verification is still typically required, so cost savings versus officer labor are limited.
Technical feasibility todayclaude-haiku-4-5-202510011/5No deployed product reliably automates prisoner head counts end-to-end; this remains a human security function with zero-tolerance error requirements. While biometric or camera systems exist, they are not in production as standalone head-count replacements in correctional facilities.
Technical feasibility todayclaude-sonnet-52/5Some facilities deploy electronic monitoring aids (RFID wristbands, camera systems), but full autonomous headcount verification without human confirmation is not standard deployed practice in corrections.

Guard facility entrances to screen visitors.

4

CI 0–9 · exposure 5 · augmentation 50 · 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 most conservative and regulation-bound sectors. Physical security gatekeeping remains a core human function with minimal automation adoption, even where technology could theoretically assist, due to liability and security protocols.
Sector adoption velocityclaude-sonnet-51/5Corrections is a highly physical, security-sensitive government sector with minimal AI-driven displacement of frontline guard duties to date.
Augmentation potentialclaude-haiku-4-5-202510013/5AI can meaningfully assist officers by automating ID document verification, cross-referencing watch-lists, and flagging risk patterns, allowing officers to focus on interpersonal assessment and decision-making. Such assistance improves efficiency without removing human judgment from the critical screening function.
Augmentation potentialclaude-sonnet-53/5AI-enhanced scanners, facial recognition, and visitor-management software can assist officers in screening and flagging risks, improving efficiency while humans remain in control.
Task automatabilityclaude-haiku-4-5-202510011/5Guarding facility entrances and screening visitors requires real-time physical presence, interpersonal judgment, and enforcement authority that cannot be automated end-to-end. While AI could assist with ID verification or risk flagging, the core task—making discretionary decisions, managing confrontations, and maintaining security posture—remains fundamentally human.
Task automatabilityclaude-sonnet-51/5This requires physical presence, real-time judgment about threats, weapons detection, and authority to detain—current AI cannot perform the physical screening and enforcement actions end-to-end.
Adoption barriersclaude-haiku-4-5-202510015/5This task has hard legal and regulatory barriers: correctional facilities are typically required by law to have licensed, trained personnel physically present at entrances with arrest and use-of-force authority. Liability, safety, and compliance regulations mandate human gatekeeping in secure facilities.
Adoption barriersclaude-sonnet-55/5Correctional facilities require sworn, authorized personnel with legal authority to detain, search, and use force—this is heavily regulated and cannot be delegated to non-human systems.
Cost vs. human wageclaude-haiku-4-5-202510012/5AI screening tools (document recognition, database lookups) exist but are modest in cost savings relative to the human officer's loaded wage, since they address only narrower sub-components and still require human oversight and decision-making on sensitive cases.
Cost vs. human wageclaude-sonnet-51/5AI sensors (metal detectors, cameras) already exist as tools but do not replace the human officer's cost; a full substitute would require robotics and legal authority far more expensive than current wages.
Technical feasibility todayclaude-haiku-4-5-202510012/5Some AI systems can assist with document verification or watch-list matching in production, but no deployed system reliably performs the full screening task autonomously. The task requires judgment calls, physical intervention capacity, and legal authority that current AI cannot independently exercise.
Technical feasibility todayclaude-sonnet-51/5No deployed product autonomously guards a correctional facility entrance and screens visitors without human officers physically present and making authorization decisions.

Serve meals, distribute commissary items, and dispense prescribed medication to prisoners.

4

CI 0–9 · exposure 5 · 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 facilities are typically government-run, resource-constrained, and resistant to automation due to security concerns and union agreements. Adoption of automation in this sector remains minimal and slow compared to commercial sectors.
Sector adoption velocityclaude-sonnet-51/5Corrections is a highly physical, security-focused, low-digitization sector with minimal AI agent deployment for direct inmate-facing physical tasks.
Augmentation potentialclaude-haiku-4-5-202510012/5AI could potentially assist with inventory management or commissary ordering, but the direct service and medication components require human presence and judgment, limiting meaningful augmentation in the core task as stated.
Augmentation potentialclaude-sonnet-52/5AI could assist with scheduling, inventory tracking, or medication administration record-keeping, but offers negligible help with the core physical act of serving meals and dispensing items.
Task automatabilityclaude-haiku-4-5-202510011/5This task requires physical presence in secure facilities, direct interaction with incarcerated individuals, and real-time decision-making around medication safety and dietary needs. Current AI and robotic systems cannot reliably handle the full physical and interpersonal complexity of this work in a correctional setting.
Task automatabilityclaude-sonnet-51/5This is a physical task requiring direct manual handling of food, commissary items, and medication distribution to incarcerated individuals in a secure environment; no current AI system can perform this physically.
Adoption barriersclaude-haiku-4-5-202510015/5Correctional facilities are highly regulated environments with strict legal requirements around medication administration and prisoner welfare. Medication must be dispensed by authorized personnel, and there are significant liability and safety accountability requirements that legally necessitate human judgment and oversight.
Adoption barriersclaude-sonnet-55/5Medication dispensing to inmates involves strict legal, medical, and security requirements including chain-of-custody, controlled substance handling, and officer authorization, creating hard regulatory and safety barriers to automation.
Cost vs. human wageclaude-haiku-4-5-202510012/5Implementing robotic or automated systems for meal service and commissary distribution requires substantial capital investment, ongoing maintenance, and integration costs that currently exceed the cost of human labor for these tasks in most correctional settings.
Cost vs. human wageclaude-sonnet-51/5AI has no viable substitute for this physical labor, so any hypothetical automation (e.g., robotics) would be far more costly than existing human labor given current technology costs and integration needs.
Technical feasibility todayclaude-haiku-4-5-202510012/5While some facilities experiment with robotic food carts and automated commissary systems, no deployed product reliably handles all three components (meal service, commissary distribution, medication dispensing) together in production. Medication dispensing in particular involves liability and verification requirements that limit current automation.
Technical feasibility todayclaude-sonnet-51/5There are no deployed AI or robotic products performing meal service, commissary distribution, or medication dispensing to prisoners in correctional facilities today; this remains entirely a human physical task.

Process or book convicted individuals into prison.

4

CI 0–7 · exposure 5 · augmentation 38 · importance 4.2/5 · click for rater detail

Panel ratings per dimension for this task
DimensionModelRatingRationale
Sector adoption velocityclaude-haiku-4-5-202510012/5Corrections is a traditional, public-sector, heavily regulated domain with slow digitization and strong legal requirements for human involvement. Adoption of AI agents in booking workflows is minimal; most jails and prisons still rely on manual or partially digitized intake processes.
Sector adoption velocityclaude-sonnet-51/5Corrections is a highly physical, low-digitization sector with minimal AI deployment for custodial tasks.
Augmentation potentialclaude-haiku-4-5-202510013/5AI can assist by pulling prior records, flagging inconsistencies in documents, and automating data entry into booking systems, but the officer remains the decision-maker. These tools provide meaningful productivity gains on routine sub-tasks while the human retains custody and legal responsibility.
Augmentation potentialclaude-sonnet-52/5AI can help with digital record-keeping, data entry, or biometric matching during booking, but it doesn't materially transform the core physical processing task.
Task automatabilityclaude-haiku-4-5-202510011/5Booking convicted individuals into prison requires human judgment, legal verification, biometric capture, physical documentation review, and authorized decision-making that cannot be automated end-to-end today. Current AI systems cannot reliably handle the variability of legal records, verify identity, or assume liability for the accuracy of intake decisions.
Task automatabilityclaude-sonnet-51/5Booking involves physical custody, searches, fingerprinting, and direct control of a person in a secure facility, none of which current AI can perform end-to-end.
Adoption barriersclaude-haiku-4-5-202510015/5Booking convicted individuals into prison is a legally mandated function requiring authorized, licensed correctional staff and chain-of-custody accountability. Liability, regulatory compliance, and the requirement for human official sign-off create hard barriers to automation.
Adoption barriersclaude-sonnet-55/5Booking requires sworn, legally authorized correctional officers with custodial authority; this is a hard legal/regulatory and security barrier.
Cost vs. human wageclaude-haiku-4-5-202510011/5AI tools that assist with booking (document management, data entry) are offset by required oversight, validation, and the skilled labor of correctional officers who must remain present for legal and safety reasons, making the total cost comparable to or higher than human-only workflows.
Cost vs. human wageclaude-sonnet-51/5There is no viable AI substitute for the physical labor and custodial authority required, so cost comparison favors the human by default.
Technical feasibility todayclaude-haiku-4-5-202510012/5While some components (data entry, record retrieval) could be assisted by AI, no deployed production system end-to-end books individuals into prison autonomously. Partial tools exist for document processing, but the full intake workflow involving custody verification, medical screening, and authorized sign-off remains human-driven.
Technical feasibility todayclaude-sonnet-51/5No deployed product performs the physical intake and custody transfer of a convicted individual; this remains entirely a human security function.

Search for and recapture escapees.

3

CI 0–5 · exposure 5 · augmentation 38 · importance 4.4/5 · 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 emphasis on human judgment and accountability; adoption of autonomous search-and-capture systems is negligible.
Sector adoption velocityclaude-sonnet-51/5Corrections is a physically-oriented, low-digitization sector with minimal AI deployment for hands-on custodial and pursuit functions.
Augmentation potentialclaude-haiku-4-5-202510013/5AI can meaningfully assist correctional officers by analyzing surveillance footage, predicting escape routes, cross-referencing fugitive records, and identifying likely locations—useful support that enhances human-led manhunts without replacing the officer's field presence and judgment.
Augmentation potentialclaude-sonnet-52/5AI can assist with tracking tools, cameras, GPS monitoring, or predictive analytics to aid search planning, but offers little help in the actual physical recapture.
Task automatabilityclaude-haiku-4-5-202510011/5Recapturing escapees requires physical apprehension, real-time threat assessment, dynamic pursuit across variable terrain, and rapid decision-making under high-risk conditions. Current AI systems cannot perform this end-to-end without human operators in the field.
Task automatabilityclaude-sonnet-51/5This is a physical, high-stakes law enforcement pursuit task requiring physical presence, use of force decisions, and real-world tracking; no AI system can perform this end-to-end.'
Adoption barriersclaude-haiku-4-5-202510015/5Capture and detention of escapees is a core government function that typically requires licensed law enforcement or correctional officers; liability, legal authority, and use-of-force regulations create hard barriers to autonomous replacement.
Adoption barriersclaude-sonnet-55/5Use of force, arrest authority, and physical apprehension are legally restricted to sworn/authorized personnel, creating hard legal and safety barriers to automation.
Cost vs. human wageclaude-haiku-4-5-202510011/5Full automation would require expensive robotics, surveillance infrastructure, and real-time processing at the scale of a fugitive manhunt—substantially more costly than directing existing correctional staff to conduct searches.
Cost vs. human wageclaude-sonnet-51/5There is no AI substitute performing this task, so cost comparison favors humans entirely; any AI cost would be additive to still-necessary human officers.
Technical feasibility todayclaude-haiku-4-5-202510012/5Deployed AI can assist with fugitive tracking (facial recognition, geolocation analysis, record matching) but cannot autonomously search or recapture; systems that handle these elements remain in pilot phases without production-scale deployment in actual escape scenarios.
Technical feasibility todayclaude-sonnet-51/5No deployed AI product searches for and physically recaptures escapees; this remains entirely a human/physical security function, at most aided by surveillance tech.

Sponsor inmate recreational activities, such as newspapers and self-help groups.

3

CI 0–5 · exposure 0 · augmentation 25 · importance 3.6/5 · click for rater detail

Panel ratings per dimension for this task
DimensionModelRatingRationale
Sector adoption velocityclaude-haiku-4-5-202510011/5Correctional facilities operate under strict regulatory frameworks requiring human staff oversight. Adoption of AI for inmate program facilitation is minimal because the task inherently requires trained human personnel responsible for inmate welfare and institutional safety.
Sector adoption velocityclaude-sonnet-51/5Corrections is a low-digitization, physically-bound sector with minimal AI agent deployment for direct inmate program administration.
Augmentation potentialclaude-haiku-4-5-202510012/5AI could minimally assist by generating activity schedules or self-help group curriculum ideas, but the core human task—building trust, managing group dynamics, and ensuring inmate safety—cannot be augmented in ways that materially improve officer productivity on this interpersonal responsibility.
Augmentation potentialclaude-sonnet-52/5AI could help draft newsletters or organize scheduling materials, but the sponsoring/supervisory aspect itself gains little from AI tools.
Task automatabilityclaude-haiku-4-5-202510011/5This task fundamentally requires human relationship-building, social judgment, and real-time interpersonal engagement. AI cannot meaningfully organize, facilitate, or sponsor actual recreational activities or groups with inmates—it lacks agency and cannot serve as a responsible organizer or monitor of inmate programs.
Task automatabilityclaude-sonnet-51/5Sponsoring and organizing inmate activities requires physical presence, supervision, relationship-building, and security judgment that AI cannot perform end-to-end.'
Adoption barriersclaude-haiku-4-5-202510015/5Sponsoring and facilitating inmate programs is a core custodial and rehabilitative responsibility that requires licensed correctional staff. Liability, duty of care, safety oversight, and regulatory requirements legally mandate human participation and sign-off.
Adoption barriersclaude-sonnet-54/5Correctional facilities require authorized staff with security clearance and institutional trust to supervise inmate programs, creating strong organizational and regulatory barriers to automation.
Cost vs. human wageclaude-haiku-4-5-202510011/5The human cost of correctional staff time is significantly lower than any attempt to deploy AI systems capable of legal accountability, supervision, and relationship-building with inmates in a correctional setting.
Cost vs. human wageclaude-sonnet-51/5There is no viable AI substitute performing this function, so cost comparison favors the human by default.
Technical feasibility todayclaude-haiku-4-5-202510011/5No deployed product performs this task. While AI can generate content or schedules, it cannot actually sponsor, authorize, or oversee inmate activities in a correctional facility—this requires human accountability and on-site presence.
Technical feasibility todayclaude-sonnet-51/5No deployed AI product organizes or sponsors inmate recreational programs; this is a custodial, in-person organizational role.

Drive passenger vehicles and trucks used to transport inmates to other institutions, courtrooms, hospitals, and work sites.

2

CI 0–4 · 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 agencies are highly conservative, budget-constrained government bodies with strong human-oversight cultures. No measurable adoption of autonomous prisoner transport is occurring in production correctional systems.
Sector adoption velocityclaude-sonnet-51/5Corrections is a slow-adopting, highly regulated, physically embodied sector with essentially no autonomous vehicle or AI transport deployment.
Augmentation potentialclaude-haiku-4-5-202510012/5GPS navigation and route optimization tools offer minimal augmentation to an officer's core task; the driving, security monitoring, and prisoner supervision responsibilities remain largely manual and human-dependent.
Augmentation potentialclaude-sonnet-52/5AI could assist with route planning, scheduling, or vehicle diagnostics, but offers minimal help with the core custodial driving and security task.
Task automatabilityclaude-haiku-4-5-202510011/5The task requires real-time navigation in dynamic environments, monitoring of potentially dangerous passengers, and immediate human judgment to respond to security threats or prisoner escape attempts. Current AI cannot reliably handle the safety-critical, unstructured nature of transporting incarcerated individuals.
Task automatabilityclaude-sonnet-51/5Physically driving vehicles while securing and monitoring inmates requires embodied presence, security judgment, and physical control that no current AI or autonomous system can perform end-to-end.
Adoption barriersclaude-haiku-4-5-202510015/5Transporting incarcerated individuals requires a licensed, accountable human officer by law in virtually all jurisdictions. Liability for escape, injury, or misconduct during transport is legally vested in named personnel, creating a hard regulatory barrier to full automation.
Adoption barriersclaude-sonnet-55/5Legal custody requirements, security protocols, and liability for inmate escape or harm mandate certified correctional officers, creating hard regulatory and safety barriers to automation.
Cost vs. human wageclaude-haiku-4-5-202510012/5Autonomous vehicles remain significantly more expensive than employing a correctional officer when accounting for vehicle cost, insurance, liability coverage, and required oversight infrastructure. The security and regulatory overhead makes AI economically uncompetitive.
Cost vs. human wageclaude-sonnet-51/5There is no viable AI substitute for this physical custodial transport task, so AI cost comparison is not applicable and effectively far more expensive/infeasible than human labor.
Technical feasibility todayclaude-haiku-4-5-202510011/5While autonomous vehicle technology exists in limited deployment, no production system reliably handles prisoner transport with the security oversight and liability management required. The combination of driving, supervision, and security responsibilities has not been demonstrated in deployed correctional systems.
Technical feasibility todayclaude-sonnet-51/5No deployed product transports and secures inmates autonomously; autonomous vehicle tech is not certified or used for correctional transport in production.

Inspect conditions of locks, window bars, grills, doors, and gates at correctional facilities to ensure security and help prevent 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 among the most conservative, rule-bound organizational environments with minimal digital transformation and strong legal requirements for human accountability in security roles; adoption of automation for critical security checks is negligible.
Sector adoption velocityclaude-sonnet-51/5Corrections is a highly physical, low-digitization sector with minimal AI agent deployment for hands-on custodial and security tasks.
Augmentation potentialclaude-haiku-4-5-202510012/5AI could provide minor assistance (e.g., image-based documentation or anomaly flagging from camera feeds), but the core task—hands-on physical inspection and security certification—requires human expertise and responsibility, limiting meaningful augmentation.
Augmentation potentialclaude-sonnet-52/5Sensors, cameras, and IoT-based monitoring systems can supplement physical checks by flagging anomalies, but they don't substitute for or meaningfully transform the manual inspection task itself.
Task automatabilityclaude-haiku-4-5-202510011/5This task requires physical presence, tactile inspection (checking locks, bars, grills), and real-time assessment of security vulnerabilities at a correctional facility. Current AI systems cannot perform these hands-on, spatially-grounded inspections without human operators physically present.
Task automatabilityclaude-sonnet-51/5This requires physical presence and manual inspection of physical security hardware in a secure facility; no current AI system can perform hands-on inspection of locks, bars, and gates.
Adoption barriersclaude-haiku-4-5-202510015/5This task has hard legal and regulatory barriers: correctional facilities are heavily regulated, security protocols typically require credentialed human staff to perform security inspections, and liability for escape or security breach would fall on the facility/operator rather than on an automated system.
Adoption barriersclaude-sonnet-55/5Correctional facility security is heavily regulated and requires certified, authorized personnel to conduct physical security checks; liability for escapes and safety failures is extremely high.
Cost vs. human wageclaude-haiku-4-5-202510011/5The cost of deploying robotic or vision-based inspection systems, plus integration, oversight, and liability insurance, would far exceed the loaded wage of a correctional officer performing routine inspections in-house.
Cost vs. human wageclaude-sonnet-51/5There is no viable AI substitute for physical inspection, so cost comparison favors the human by default since AI cannot perform the task at all.
Technical feasibility todayclaude-haiku-4-5-202510011/5No deployed product reliably performs autonomous security inspections of physical correctional infrastructure today. While computer vision can analyze images, it cannot autonomously traverse facilities and conduct the tactile, compliance-critical inspections this task demands.
Technical feasibility todayclaude-sonnet-51/5No deployed product performs physical security hardware inspection in correctional facilities; this remains a manual task requiring a human physically present.

Search prisoners and vehicles and conduct shakedowns of cells for valuables and contraband, such as weapons or drugs.

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 traditionally low-digitization, risk-averse organizations with strong labor unions and legal constraints. Adoption of automation for core custody tasks like searches is minimal and unlikely in the foreseeable future.
Sector adoption velocityclaude-sonnet-51/5Corrections is a low-digitization, physically-oriented sector with minimal AI agent deployment for hands-on custodial functions.
Augmentation potentialclaude-haiku-4-5-202510012/5Limited augmentation potential exists; imaging or detection tools might assist in narrowing search areas, but the core task—physical inspection and judgment about contraband—remains overwhelmingly human-dependent. AI tools offer marginal productivity gains at best.
Augmentation potentialclaude-sonnet-52/5AI-enabled tools like metal detectors, X-ray scanners, or drug-sniffing sensor systems can assist by flagging contraband, but they only marginally support the human-led physical search process.
Task automatabilityclaude-haiku-4-5-202510011/5This task requires physical inspection, spatial reasoning, and physical handling that current AI systems cannot perform. Robots exist for narrow purposes but cannot reliably execute the full end-to-end task of searching bodies, vehicles, and cells while identifying contraband with the judgment and safety standards required.
Task automatabilityclaude-sonnet-51/5This is a physical, hands-on security task requiring physical searches of people, vehicles, and cells that current AI cannot perform end-to-end; no robotic or AI system can conduct physical pat-downs or cell shakedowns.
Adoption barriersclaude-haiku-4-5-202510015/5This task is firmly bound by legal and regulatory requirements: only authorized personnel can conduct searches, and liability for improper search procedures (damage to property, injury, rights violations) creates strong legal barriers. Human contact and authority are non-negotiable elements.
Adoption barriersclaude-sonnet-55/5Legal, security, and custodial authority require sworn, licensed correctional officers to conduct searches and use of force; this is a highly regulated, human-contact-mandated function with major liability and chain-of-custody implications.
Cost vs. human wageclaude-haiku-4-5-202510011/5Any technological solution (scanners, robots, imaging) would be prohibitively expensive compared to the cost of correctional officer labor for this task, especially given the low-volume, high-variability nature of searches across diverse facility settings.
Cost vs. human wageclaude-sonnet-51/5There is no viable AI substitute performing the physical search itself, so any AI cost comparison is moot—human officers remain the only means of executing this task.
Technical feasibility todayclaude-haiku-4-5-202510011/5No deployed AI system reliably performs prisoner searches, vehicle inspections, or cell shakedowns independently. While imaging and detection technologies exist, they require human interpretation and are not deployed as autonomous systems in correctional facilities.
Technical feasibility todayclaude-sonnet-51/5No deployed product performs physical searches of inmates or contraband sweeps; sensor-based contraband detection tools exist only as narrow aids, not autonomous search systems.

Use weapons, handcuffs, and physical force to maintain discipline and order among prisoners.

0

CI 0–0 · exposure 0 · augmentation 13 · 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 remain low-digitization, human-intensive environments with strong institutional and legal constraints against automation of security and control functions. Adoption of AI for this specific task is negligible.
Sector adoption velocityclaude-sonnet-51/5Corrections is a low-digitization, physical-labor sector with minimal AI deployment for direct inmate control functions.
Augmentation potentialclaude-haiku-4-5-202510011/5While surveillance or alert systems might assist officers in monitoring, AI offers no meaningful assistance with the core task of using physical force, weapons, or handcuffs to maintain discipline—these require direct human execution.
Augmentation potentialclaude-sonnet-52/5AI can support situational awareness via surveillance analytics or risk-flagging, but offers negligible assistance during the actual physical act of restraint or force application.
Task automatabilityclaude-haiku-4-5-202510011/5Physical force, weapon handling, and real-time disciplinary decisions in volatile human situations require embodied presence, judgment under threat, and split-second contextual reasoning that no current AI system can perform. Automation of this task is not technically feasible today.
Task automatabilityclaude-sonnet-51/5This is a physical, safety-critical task requiring embodied human judgment, force application, and legal authority; no AI system can perform physical restraint or weapon use.
Adoption barriersclaude-haiku-4-5-202510015/5Hard legal and liability barriers: correctional officers must be licensed, trained peace officers with statutory authority to use force and carry weapons. Prisoners have rights protections that require human judgment and accountability. Only certified humans can legally execute this role.
Adoption barriersclaude-sonnet-55/5Use of force and restraint by correctional officers is tightly regulated, requires sworn/authorized personnel, and carries significant legal and liability consequences, making substitution legally impossible today.
Cost vs. human wageclaude-haiku-4-5-202510011/5AI cannot perform this task at all, making direct cost comparison meaningless. A human correctional officer is the only option.
Cost vs. human wageclaude-sonnet-51/5There is no AI substitute performing this physical enforcement task, so cost comparison is inapplicable/AI is not a viable cheaper alternative.
Technical feasibility todayclaude-haiku-4-5-202510011/5No deployed AI product performs physical restraint, weapon deployment, or direct prisoner control. This task fundamentally requires a physical human agent with legal authority and real-time decision-making in unpredictable, high-risk environments.
Technical feasibility todayclaude-sonnet-51/5No deployed product performs physical control of inmates; this remains entirely a human physical and legal responsibility with no robotics substitute in production.

Supervise and coordinate work of other correctional service officers.

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 facilities are highly regulated, traditionalist institutions with strong hierarchical and human-accountability structures. Supervision automation is not on any sector roadmap, and cultural and legal barriers to removing human supervisors are extremely high.
Sector adoption velocityclaude-sonnet-51/5Corrections is a low-digitization, physically-grounded sector with minimal AI agent deployment in operational supervisory roles.
Augmentation potentialclaude-haiku-4-5-202510012/5AI might assist with scheduling, roster analytics, or incident logging, but the core act—real-time supervision, personnel direction, and duty-of-care decisions—remains fundamentally human. Marginal assistance does not substantially transform supervisory productivity.
Augmentation potentialclaude-sonnet-52/5AI scheduling or reporting tools may support administrative aspects of coordination, but the core supervisory judgment and interpersonal authority receive little meaningful AI assistance.
Task automatabilityclaude-haiku-4-5-202510011/5Supervising and coordinating work of other officers requires dynamic interpersonal judgment, conflict resolution, real-time decision-making under stress, and accountability for human safety—core human responsibilities that current AI cannot perform end-to-end. No system can replace the duty-of-care and legal authority a supervisor must exercise.
Task automatabilityclaude-sonnet-51/5Supervising and coordinating human officers in a secure custodial environment requires physical presence, real-time judgment, and authority that current AI cannot replicate or execute end-to-end.
Adoption barriersclaude-haiku-4-5-202510015/5Correctional supervision is a licensed, regulated position with explicit statutory authority to direct and discipline staff, ensure facility safety, and make binding personnel decisions. Legal and organizational frameworks explicitly require a human supervisor with accountability and decision-making authority.
Adoption barriersclaude-sonnet-55/5Supervisory authority in correctional settings is tied to sworn/certified officer status, chain-of-command accountability, and legal liability, making this a hard institutional and regulatory barrier.
Cost vs. human wageclaude-haiku-4-5-202510011/5An AI system would need to match or exceed the full loaded cost of a supervisory officer (salary, benefits, training, liability) while handling live staff coordination, performance reviews, and decision-making—a threshold not approached by current AI infrastructure.
Cost vs. human wageclaude-sonnet-51/5There is no viable AI substitute performing this supervisory role, so cost comparison favors the human by default.
Technical feasibility todayclaude-haiku-4-5-202510011/5No deployed product today reliably supervises staff, manages personnel discipline, handles grievances, or coordinates live operational teams in correctional settings. This task requires human judgment, legal authority, and presence that production AI systems do not possess.
Technical feasibility todayclaude-sonnet-51/5No deployed product performs supervisory coordination of correctional staff; this remains purely a human management function in facilities today.

Take prisoners into custody and escort to locations within and outside of facility, such as visiting room, courtroom, or airport.

0

CI 0–0 · exposure 0 · augmentation 13 · importance 4.1/5 · click for rater detail

Panel ratings per dimension for this task
DimensionModelRatingRationale
Sector adoption velocityclaude-haiku-4-5-202510011/5Correctional facilities are highly regulated, physically bound organizations with no measurable adoption of autonomous prisoner transport systems. Adoption velocity is near zero.
Sector adoption velocityclaude-sonnet-51/5Corrections is a highly physical, low-digitization sector with essentially no automation of custodial escort functions; adoption of AI for this specific task is nonexistent.
Augmentation potentialclaude-haiku-4-5-202510011/5Current AI offers minimal assistance for the core task of physically escorting and maintaining custody of prisoners; body cameras or logistics software offer only peripheral support, not transformation of officer productivity in the custody function.
Augmentation potentialclaude-sonnet-52/5AI can support scheduling, tracking, or monitoring systems (e.g., locating logs, alerts) but offers minimal direct assistance to the physical act of escorting a prisoner.
Task automatabilityclaude-haiku-4-5-202510011/5This task requires physical presence, custody control, and real-time responsiveness to security threats. Current AI systems cannot physically restrain, escort, or assume legal responsibility for prisoner custody and safety.
Task automatabilityclaude-sonnet-51/5This requires physical presence, physical control of a person, and use-of-force judgment that no current AI or robotic system can perform; it is fundamentally a physical custody task.
Adoption barriersclaude-haiku-4-5-202510015/5Strong legal and regulatory barriers exist: correctional officers are licensed positions with statutory duties, and liability for prisoner safety falls on authorized human officers. Laws explicitly require human custody and supervision of detained individuals.
Adoption barriersclaude-sonnet-55/5Legal custody, use of force, and security authority are strictly reserved for sworn/certified correctional officers under statute and institutional policy, making substitution legally impossible.
Cost vs. human wageclaude-haiku-4-5-202510011/5The cost of a robotic system capable of secure prisoner transport and supervision, plus infrastructure and maintenance, far exceeds the loaded wage of a correctional officer.
Cost vs. human wageclaude-sonnet-51/5There is no AI system that performs this physical task, so no meaningful cost comparison exists; a human officer is the only viable option.
Technical feasibility todayclaude-haiku-4-5-202510011/5No deployed AI system can perform the physical and custodial aspects of this task. Robotics for prisoner transport exist only in research and are far from reliable deployment in real correctional facilities.
Technical feasibility todayclaude-sonnet-51/5No deployed product exists that physically escorts or takes custody of incarcerated individuals; this remains entirely a human security function.

Settle disputes between inmates.

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 facilities operate in highly regulated, physically constrained environments with legal accountability requirements and low digitization rates for core security functions. Adoption of AI for inmate management remains minimal.
Sector adoption velocityclaude-sonnet-51/5Corrections is a highly physical, low-digitization sector with minimal AI deployment for direct inmate interaction or conflict management.
Augmentation potentialclaude-haiku-4-5-202510012/5AI might provide limited assistance through sentiment analysis of communication or alerting systems for escalation risk, but the core task of settling disputes requires human presence, authority, and judgment that AI cannot meaningfully augment today.
Augmentation potentialclaude-sonnet-52/5AI can support with surveillance analytics or flagging tension via monitoring systems, but it provides little direct assistance to the officer during real-time dispute resolution.
Task automatabilityclaude-haiku-4-5-202510011/5Settling disputes between inmates requires nuanced understanding of social context, interpersonal dynamics, threat assessment, and real-time judgment in volatile situations. Current AI systems cannot reliably perform the core safety-critical and judgment-heavy elements of this task.
Task automatabilityclaude-sonnet-51/5Settling disputes between inmates requires physical presence, real-time judgment, authority, and de-escalation of potentially violent situations that no current AI system can perform end-to-end.
Adoption barriersclaude-haiku-4-5-202510015/5Correctional officers are state or federal employees with legal authority and accountability; settling disputes is a core function requiring human judgment, presence, and legal liability. Liability asymmetry and regulatory requirements strongly protect this role from automation.
Adoption barriersclaude-sonnet-55/5Only sworn/authorized correctional staff can legally exercise custodial authority, use force if needed, and be held liable for maintaining order, making this a hard institutional and legal barrier to automation.
Cost vs. human wageclaude-haiku-4-5-202510011/5AI would require significant human oversight, monitoring, and intervention capability to be viable, making it more expensive than simply employing trained correctional staff to handle disputes directly.
Cost vs. human wageclaude-sonnet-51/5There is no AI substitute performing this function, so cost comparison favors the human entirely; any AI cost would be additive at best (e.g., monitoring tools) rather than substitutive.
Technical feasibility todayclaude-haiku-4-5-202510011/5No deployed AI product performs inmate dispute resolution in production correctional facilities. This task requires physical presence, immediate authority, de-escalation expertise, and legal accountability—none of which current AI systems can reliably provide.
Technical feasibility todayclaude-sonnet-51/5No deployed AI product mediates or resolves in-person conflicts between incarcerated individuals; this remains entirely a human function requiring physical intervention and situational authority.

Counsel inmates and respond to legitimate questions, concerns, and requests.

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 systems are highly regulated, budget-constrained, and institutionally conservative. There is minimal adoption of AI for inmate counseling or response functions, and legal/union resistance would be substantial.
Sector adoption velocityclaude-sonnet-51/5Corrections is a highly physical, low-digitization, security-constrained sector with minimal AI agent deployment for direct inmate interaction.
Augmentation potentialclaude-haiku-4-5-202510012/5While AI could potentially assist with documentation, flagging urgent keywords in requests, or suggesting resources, it offers minimal meaningful augmentation for the core counseling and judgment-intensive aspects of responding to inmate concerns.
Augmentation potentialclaude-sonnet-52/5AI could help officers with reference information, documentation, or flagging concerning patterns in requests, but it offers little direct assistance in the interpersonal counseling moment itself.
Task automatabilityclaude-haiku-4-5-202510011/5This task requires nuanced emotional intelligence, discretionary judgment about inmate welfare, and the ability to identify and de-escalate sensitive interpersonal situations—capabilities that current AI cannot reliably perform. The task fundamentally depends on human presence, trust-building, and contextual decision-making that AI cannot replicate at equal quality.
Task automatabilityclaude-sonnet-51/5This requires real-time physical presence, situational judgment, and trust-building with a vulnerable, sometimes volatile population inside a secure facility; no AI system can perform this end-to-end today.'
Adoption barriersclaude-haiku-4-5-202510015/5Correctional facilities operate under strict legal frameworks where human officers are required by statute to conduct welfare checks, respond to grievances, and provide counseling oversight. Liability for inmate safety and duty-of-care requirements create hard barriers to full automation.
Adoption barriersclaude-sonnet-55/5Correctional officer duties are governed by strict legal, custodial, and safety/security requirements mandating trained, authorized personnel physically present with inmates, and liability for mishandling is severe.
Cost vs. human wageclaude-haiku-4-5-202510011/5Any AI system would require full human oversight and verification of responses, adding cost rather than reducing it; the human officer must remain present and accountable, so automation yields no cost savings.
Cost vs. human wageclaude-sonnet-51/5Since no viable AI substitute exists for this in-person, security-sensitive interaction, the human is the only functional option, making AI not cheaper but simply inapplicable.
Technical feasibility todayclaude-haiku-4-5-202510011/5No production system can reliably counsel inmates or handle sensitive grievances and mental health concerns autonomously. This task requires licensed human judgment and legal accountability that AI systems cannot assume.
Technical feasibility todayclaude-sonnet-51/5There are no deployed products that counsel inmates or handle their concerns in correctional settings; this remains outside current product scope.

Investigate crimes that have occurred within an institution, or assist police in their investigations of crimes and inmates.

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 institutions operate in heavily regulated, lower-digitization environments with strong adherence to established investigative protocols; AI adoption in investigation is minimal and limited to optional document assistance.
Sector adoption velocityclaude-sonnet-51/5Corrections is a physically intensive, low-digitization government sector with minimal AI agent deployment for investigative work; adoption here is essentially nonexistent.
Augmentation potentialclaude-haiku-4-5-202510012/5AI could modestly assist by flagging patterns in data or organizing evidence, but most of the investigative task—interviews, judgment, chain of custody, legal reasoning—remains outside current augmentation scope and requires human authority.
Augmentation potentialclaude-sonnet-52/5AI can help with report drafting, searching records, or flagging patterns in surveillance data, but it offers only marginal assistance to the core investigative task of interviewing and evidence-gathering in a secure facility.
Task automatabilityclaude-haiku-4-5-202510011/5Investigating crimes requires complex judgment, witness interviews, evidence evaluation, and understanding of institutional context and human intent—capabilities that current AI systems cannot perform end-to-end with reliability or legal defensibility today.
Task automatabilityclaude-sonnet-51/5Investigating crimes within a secure institution requires physical presence, interviewing inmates and staff, gathering physical evidence, and exercising judgment in a high-stakes security context that current AI cannot perform end-to-end.
Adoption barriersclaude-haiku-4-5-202510015/5Criminal investigation requires legally authorized personnel; only trained investigators can conduct interviews, make charges, and testify in court—these are hard legal and institutional barriers that prevent substitution regardless of AI capability.
Adoption barriersclaude-sonnet-55/5Criminal investigations within correctional facilities involve chain-of-custody, legal authority, safety/security clearances, and evidentiary standards that require sworn, authorized personnel—hard legal and institutional barriers to automation.
Cost vs. human wageclaude-haiku-4-5-202510011/5The cost of AI systems, training, integration, and human oversight required to partially assist investigation still substantially exceeds the operational benefit relative to trained human investigators carrying out this work.
Cost vs. human wageclaude-sonnet-51/5There is no AI system performing this task, so no cost comparison favors AI; the human officer remains the only viable performer at any cost.
Technical feasibility todayclaude-haiku-4-5-202510011/5No deployed product reliably conducts criminal investigations autonomously; AI tools may assist with document review or pattern flagging, but investigative authority and accountability remain inherently human functions with no production automation at this task level.
Technical feasibility todayclaude-sonnet-51/5No deployed product conducts institutional criminal investigations autonomously; this remains firmly a human investigative function with no production AI substitute.

Assign duties to inmates, providing instructions as needed.

0

CI 0–0 · exposure 0 · augmentation 25 · importance 3.9/5 · click for rater detail

Panel ratings per dimension for this task
DimensionModelRatingRationale
Sector adoption velocityclaude-haiku-4-5-202510011/5Correctional facilities are typically lower-digitization organizations with conservative adoption patterns and deep institutional resistance to automation in functions involving custody, safety, and inmate management.
Sector adoption velocityclaude-sonnet-51/5Corrections is a highly institutionalized, low-digitization, security-sensitive sector with minimal AI deployment in direct inmate supervision tasks.
Augmentation potentialclaude-haiku-4-5-202510012/5AI could potentially assist with routine administrative tracking or duty-schedule drafting, but the core task—communicating with, assessing, and directing inmates—offers limited augmentation potential because human authority and presence are irreplaceable.
Augmentation potentialclaude-sonnet-52/5AI could help generate duty schedules or track task assignments in software systems, but the interpersonal instruction and oversight component sees little AI assistance today.
Task automatabilityclaude-haiku-4-5-202510011/5This task requires real-time interpersonal communication with individuals who may be hostile or uncooperative, ongoing judgment about inmate capability and behavior, and immediate responsiveness to changing conditions—core elements of human authority and discretion that current AI systems cannot reliably execute end-to-end.
Task automatabilityclaude-sonnet-51/5This task requires physical presence, real-time authority, and situational judgment within a secure facility that current AI systems cannot perform end-to-end.
Adoption barriersclaude-haiku-4-5-202510015/5Correctional officers operate under strict legal and supervisory authority structures; there are hard barriers rooted in liability law, criminal justice regulation, and the requirement that a licensed/authorized human retain decision-making power over inmate assignments and safety.
Adoption barriersclaude-sonnet-55/5Correctional officers are legally authorized, sworn personnel bound by institutional policy, security regulations, and liability concerns that require a human to directly manage and supervise inmates.
Cost vs. human wageclaude-haiku-4-5-202510011/5The overhead of AI systems, integration, and mandatory human oversight would exceed the cost of direct human assignment, given the safety and liability demands of the correctional setting.
Cost vs. human wageclaude-sonnet-51/5There is no viable AI substitute performing this function, so cost comparison favors the human officer by default.
Technical feasibility todayclaude-haiku-4-5-202510011/5No deployed AI system currently performs the live assignment and instruction of duties to incarcerated persons; this would require autonomous agents operating in high-stakes, safety-critical environments where human accountability and judgment are non-negotiable.
Technical feasibility todayclaude-sonnet-51/5No deployed product assigns and supervises inmate work duties; this remains entirely a human custodial function.

Issue clothing, tools, and other authorized items to inmates.

0

CI 0–0 · exposure 0 · augmentation 25 · importance 3.6/5 · click for rater detail

Panel ratings per dimension for this task
DimensionModelRatingRationale
Sector adoption velocityclaude-haiku-4-5-202510011/5Correctional facilities are government/public institutions with high regulatory burden, legacy systems, and limited technical adoption resources. Automation of inmate-facing operations faces institutional resistance and is not occurring in practice.
Sector adoption velocityclaude-sonnet-51/5Corrections is a low-digitization, physically-bound sector with minimal AI adoption for direct custodial tasks.
Augmentation potentialclaude-haiku-4-5-202510012/5AI could potentially assist with inventory tracking or record-keeping through barcode/RFID systems, but the core task of physically issuing and authorizing items to inmates fundamentally requires human judgment, identity verification, and accountability that AI cannot augment meaningfully.
Augmentation potentialclaude-sonnet-52/5Inventory management or tracking software could mildly assist in logging issued items, but it does not meaningfully transform the physical task itself.
Task automatabilityclaude-haiku-4-5-202510011/5Issuing items to inmates requires physical handling and distribution of goods, as well as real-time inventory verification against inmate records and authorization lists. Current AI cannot physically retrieve, verify identity, or distribute items to specific individuals in a secure facility setting.
Task automatabilityclaude-sonnet-51/5This is a physical task requiring in-person handling, inventory control, and secure custody transfer of physical items to inmates; no AI system performs the physical distribution itself.
Adoption barriersclaude-haiku-4-5-202510015/5Correctional facilities are highly regulated with strict legal requirements, security protocols, and liability standards. Only authorized human staff are permitted to manage inmate property and access to items, with extensive oversight and accountability mandates that cannot be delegated to automated systems.
Adoption barriersclaude-sonnet-55/5Correctional facilities require sworn, authorized personnel to handle security-sensitive inmate interactions and custody of items, creating strict legal and institutional barriers to automation.
Cost vs. human wageclaude-haiku-4-5-202510011/5The infrastructure, robotics, computer vision, and integration costs required to automate item issuance far exceed the labor cost of a correctional officer performing this task manually.
Cost vs. human wageclaude-sonnet-51/5AI has no capability to physically perform this task, so any comparison favors the human officer who must be present regardless of software costs.
Technical feasibility todayclaude-haiku-4-5-202510011/5No deployed AI system can autonomously perform end-to-end inventory issuance to inmates, including physical distribution, identity verification, and security compliance checks in correctional environments.
Technical feasibility todayclaude-sonnet-51/5No deployed AI product physically issues items to inmates; at most, inventory tracking software exists but the human physical action remains the core task.

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.