First-Line Supervisors of Police and Detectives
33-1012.00Median wage $106,040/yr154,610 employed (US)Rank #727 of 923 scored · top 79% by substitution
Directly supervise and coordinate activities of members of police force.
Sub-scores
0–100 · band = confidence interval from rater disagreement
Substitution18
Exposure19
Augmentation49
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
20 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%18
panel mean rating 1.7/5 → substitution pressure 18/100
Technical feasibility todayw 20%20
panel mean rating 1.8/5 → substitution pressure 20/100
Cost vs. human wagew 15%19
panel mean rating 1.8/5 → substitution pressure 19/100
Adoption barriersw 20%inverted — strong barriers lower the score18
panel mean rating 4.3/5 (barrier strength) → substitution pressure 18/100
Sector adoption velocityw 10%14
panel mean rating 1.5/5 → substitution pressure 14/100
Task breakdown (20 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.
Requisition and issue equipment and supplies.
61CI 51–70 · exposure 62 · augmentation 75 · importance 3.4/5 · click for rater detail
Requisition and issue equipment and supplies.
61| Dimension | Model | Rating | Rationale |
|---|---|---|---|
| Sector adoption velocity | claude-haiku-4-5-20251001 | 2/5 | Law enforcement is a laggard sector for digital transformation relative to information and finance sectors. Most police departments still rely on manual or semi-automated legacy systems; AI-assisted requisitioning is not yet mainstream in production across departments. |
| Sector adoption velocity | claude-sonnet-5 | 2/5 | Law enforcement agencies are generally slower adopters of AI-driven administrative systems compared to private-sector information/finance industries, though basic inventory software is common. |
| Augmentation potential | claude-haiku-4-5-20251001 | 4/5 | AI can significantly assist supervisors by auto-populating requisition forms, flagging inventory shortages, recommending standard equipment allocations, and organizing approvals, substantially reducing paperwork and cognitive load while the supervisor retains decision authority. |
| Augmentation potential | claude-sonnet-5 | 4/5 | AI-enabled inventory and procurement tools can significantly streamline tracking, forecasting needs, and automating reorder alerts, greatly aiding the supervisor while they retain final decision authority. |
| Task automatability | claude-haiku-4-5-20251001 | 4/5 | AI systems can automate most of the workflow: inventory tracking, automated reorder triggers, requisition form completion, and equipment allocation logic. However, final approval and authorization by a human supervisor typically remains required for compliance and accountability reasons, preventing a full end-to-end autonomous solution that achieves the ≥50% time-saving bar without human sign-off. |
| Task automatability | claude-sonnet-5 | 3/5 | Requisitioning and issuing equipment involves inventory tracking, ordering, and record-keeping that can largely be handled by software systems, but physical issuance and approval judgment still require human involvement., so only partial end-to-end automation is feasible. |
| Adoption barriers | claude-haiku-4-5-20251001 | 3/5 | Police departments operate under civil-service rules, procurement regulations, and accountability frameworks that mandate human authorization and sign-off on equipment issues. While AI can streamline the work, a supervisor or authorized human must approve requisitions, creating meaningful friction to full automation. |
| Adoption barriers | claude-sonnet-5 | 2/5 | No licensing requirement mandates a sworn officer perform requisitioning, but departmental procurement rules, budget authorization chains, and accountability for controlled equipment (weapons, uniforms) create moderate organizational friction. |
| Cost vs. human wage | claude-haiku-4-5-20251001 | 4/5 | AI-driven inventory and requisition systems have low per-transaction costs compared to supervisor labor (wage-loaded) to manually review, approve, and process each equipment request. Once deployed, marginal cost per requisition is minimal. |
| Cost vs. human wage | claude-sonnet-5 | 4/5 | Automated inventory/procurement systems are inexpensive to run compared to a supervisor's time spent manually tracking and ordering supplies, though some human oversight and physical handling remain necessary. |
| Technical feasibility today | claude-haiku-4-5-20251001 | 3/5 | Enterprise inventory management and procurement software with AI integration exists and performs well in corporate settings, but police departments often use legacy systems with limited automation. Deployed AI-assisted requisition systems exist but are not yet standard in law enforcement, making this only partially mature in this sector. |
| Technical feasibility today | claude-sonnet-5 | 4/5 | Inventory and procurement management software is mature and widely deployed in police departments and other organizations, reliably tracking stock, generating purchase orders, and flagging reorder needs. |
Prepare work schedules and assign duties to subordinates.
51CI 48–54 · exposure 50 · augmentation 75 · importance 4.2/5 · click for rater detail
Prepare work schedules and assign duties to subordinates.
51| Dimension | Model | Rating | Rationale |
|---|---|---|---|
| Sector adoption velocity | claude-haiku-4-5-20251001 | 3/5 | Many police departments use scheduling software, but adoption is uneven—larger agencies move faster, small and rural departments lag. Shift to AI-driven optimization is in pilot phase, not yet standard practice across the sector. |
| Sector adoption velocity | claude-sonnet-5 | 2/5 | Law enforcement agencies are typically slow, budget-constrained public sector adopters of new scheduling technology compared to fast-moving private sectors like finance or tech. |
| Augmentation potential | claude-haiku-4-5-20251001 | 4/5 | AI scheduling assistants already boost supervisor productivity by generating compliant options, flagging conflicts, and handling rotation logic. Supervisors retain final authority and can iteratively refine suggestions, making this a strong human-in-the-loop augmentation case. |
| Augmentation potential | claude-sonnet-5 | 4/5 | AI-assisted scheduling tools can meaningfully speed up draft schedule creation and flag conflicts, letting supervisors focus on exceptions and personnel judgment calls. |
| Task automatability | claude-haiku-4-5-20251001 | 3/5 | AI can assist in generating candidate schedules using constraints and optimization, but must handle complex human factors (availability, preferences, union rules, coverage requirements, fairness). Current systems can automate 40-60% of the scheduling process with human review and adjustment required for edge cases and personnel considerations. |
| Task automatability | claude-sonnet-5 | 3/5 | Scheduling optimization and duty assignment based on constraints (coverage needs, seniority rules, availability) is a well-structured problem AI can substantially assist with, though final judgment on personnel-specific factors and unpredictable operational needs still requires human oversight. |
| Adoption barriers | claude-haiku-4-5-20251001 | 3/5 | Civil service rules, union agreements, and labor law create friction. While not absolute legal barriers to automation, contractual protections and collective bargaining agreements mean adoption requires negotiation and remains subject to grievance processes, moderating substitution velocity. |
| Adoption barriers | claude-sonnet-5 | 3/5 | No licensing requirement mandates a human create schedules, but union contracts, seniority rules, and supervisor accountability for shift coverage create moderate organizational friction against full automation. |
| Cost vs. human wage | claude-haiku-4-5-20251001 | 4/5 | Modern workforce scheduling software is relatively inexpensive per shift scheduled compared to supervisor time spent manually building schedules. The cost ratio favors automation by several multiples, especially as agencies already invest in dispatch and management systems. |
| Cost vs. human wage | claude-sonnet-5 | 3/5 | Scheduling software licenses plus configuration and human review costs are comparable to the marginal supervisory time saved; savings exist but aren't order-of-magnitude given integration and oversight needs. |
| Technical feasibility today | claude-haiku-4-5-20251001 | 3/5 | Scheduling software exists in production (workforce management tools used by many agencies), but most require substantial human oversight and customization. AI-driven scheduling assistants exist but are not yet handling the full complexity of police shift logistics autonomously at scale. |
| Technical feasibility today | claude-sonnet-5 | 3/5 | Commercial workforce scheduling software with AI-driven optimization is deployed in public safety and shift-work contexts, but police-specific scheduling involves union rules, seniority, and situational judgment that limit full automation in production today. |
Maintain logs, prepare reports, and direct the preparation, handling, and maintenance of departmental records.
41CI 39–43 · exposure 50 · augmentation 63 · importance 4.1/5 · click for rater detail
Maintain logs, prepare reports, and direct the preparation, handling, and maintenance of departmental records.
41| Dimension | Model | Rating | Rationale |
|---|---|---|---|
| Sector adoption velocity | claude-haiku-4-5-20251001 | 2/5 | Police departments have been slow to adopt AI-driven record automation due to risk aversion, budget constraints, union concerns, and the highly regulated nature of evidence and records handling; adoption remains pilot-stage in most jurisdictions. |
| Sector adoption velocity | claude-sonnet-5 | 2/5 | Law enforcement agencies are historically slow to adopt new technology due to budget constraints, unionized workforces, and regulatory/procurement hurdles. |
| Augmentation potential | claude-haiku-4-5-20251001 | 3/5 | AI can assist supervisors by auto-populating routine fields, drafting report templates, and organizing logs, meaningfully improving productivity while the supervisor retains responsibility for accuracy, legal compliance, and final sign-off. |
| Augmentation potential | claude-sonnet-5 | 4/5 | AI can meaningfully assist with drafting, summarizing, and organizing reports and logs, improving supervisor efficiency while they retain oversight and final responsibility. |
| Task automatability | claude-haiku-4-5-20251001 | 3/5 | Record management, log entry standardization, and basic report generation can be partially automated with current AI systems; however, the supervisory oversight, discretionary decision-making about record classification, and handling of sensitive information require human judgment, limiting full automation to roughly half the workload. |
| Task automatability | claude-sonnet-5 | 3/5 | Log maintenance and standard report drafting from structured inputs can be automated with AI-assisted templates and records-management software, but 'directing' staff and ensuring compliance with legal/evidentiary standards requires human judgment. |
| Adoption barriers | claude-haiku-4-5-20251001 | 4/5 | Police records are subject to strict regulatory requirements (chain of custody, evidence handling, public records laws, and audit trails), and legal liability for record accuracy or improper handling creates strong organizational and compliance barriers to full automation. |
| Adoption barriers | claude-sonnet-5 | 4/5 | Police records often have legal evidentiary status, chain-of-custody and public-records requirements, and supervisory sign-off obligations that limit full automation. |
| Cost vs. human wage | claude-haiku-4-5-20251001 | 2/5 | While AI can reduce data entry and transcription costs, the supervisory oversight, human review required for accuracy and legal compliance, and integration with existing departmental systems mean the all-in cost remains comparable to or potentially higher than a human performing the task. |
| Cost vs. human wage | claude-sonnet-5 | 3/5 | AI can cut time spent on routine documentation, but integration with legacy police records systems and required human review keeps costs only moderately below fully manual processes. |
| Technical feasibility today | claude-haiku-4-5-20251001 | 3/5 | Document management systems with AI-assisted logging and report drafting exist and are deployed in some police departments, but error rates in records handling, liability concerns around accuracy, and the need for human review of sensitive content mean current products have material limitations rather than full reliability. |
| Technical feasibility today | claude-sonnet-5 | 3/5 | Records management systems and AI drafting tools exist and are used in some departments, but adoption is uneven and error-checking/oversight is still required for legal accuracy. |
Inform personnel of changes in regulations and policies, implications of new or amended laws, and new techniques of police work.
36CI 25–48 · exposure 38 · augmentation 63 · importance 4.1/5 · click for rater detail
Inform personnel of changes in regulations and policies, implications of new or amended laws, and new techniques of police work.
36| Dimension | Model | Rating | Rationale |
|---|---|---|---|
| Sector adoption velocity | claude-haiku-4-5-20251001 | 2/5 | Law enforcement remains a traditionally hierarchical, human-communication-oriented sector with slow digital transformation relative to tech or finance. While some departments use learning management systems, autonomous or near-autonomous policy briefing is not yet a measured adoption pattern in the sector. |
| Sector adoption velocity | claude-sonnet-5 | 2/5 | Law enforcement is a traditionally slow-adopting, hierarchical, and under-digitized sector, with AI tool adoption for internal policy communication still nascent and pilot-stage at best. |
| Augmentation potential | claude-haiku-4-5-20251001 | 3/5 | AI can assist by summarizing new regulations, flagging key changes, and drafting initial policy briefs that the supervisor then personalizes and delivers. This augmentation raises supervisory productivity on the information-processing side, though the final communication remains human-centric. |
| Augmentation potential | claude-sonnet-5 | 4/5 | AI can meaningfully help supervisors by drafting summaries of legal changes, creating training materials, and answering officer questions, substantially speeding up preparation and dissemination while the supervisor remains responsible for accuracy and delivery. |
| Task automatability | claude-haiku-4-5-20251001 | 2/5 | While AI can draft summaries of regulatory changes and policy updates, the task fundamentally requires interpreting implications for frontline officers and communicating context-specific guidance—activities that demand human judgment and authority. Current AI cannot reliably handle the nuanced, situational communication that differentiates this supervisory duty, and no automation achieves the required quality parity. |
| Task automatability | claude-sonnet-5 | 3/5 | AI can draft, summarize, and disseminate policy/legal updates and training content, but confirming understanding, answering nuanced questions, and ensuring accountability for compliance still require human supervisory involvement. |
| Adoption barriers | claude-haiku-4-5-20251001 | 4/5 | Police supervisors bear legal and organizational responsibility for ensuring personnel understand regulations and implications; this accountability and sign-off requirement creates a structural barrier. Departments and unions also expect human supervisory communication, and liability for miscommunicated policy falls on the supervisor personally. |
| Adoption barriers | claude-sonnet-5 | 3/5 | No explicit licensing requirement blocks AI assistance, but chain-of-command accountability and legal liability for miscommunicating regulations create moderate organizational friction against full automation. |
| Cost vs. human wage | claude-haiku-4-5-20251001 | 2/5 | AI tools for document processing and draft generation have low direct costs, but the supervisor's wage remains the controlling cost since the human must review, contextualize, and deliver the briefing. Overall cost savings are minimal given that oversight and adaptation still dominate. |
| Cost vs. human wage | claude-sonnet-5 | 3/5 | AI-generated summaries and communications are cheap to produce, but the overall task also includes verification, accountability, and personal communication with officers, keeping total cost comparable to a supervisor's time rather than dramatically cheaper. |
| Technical feasibility today | claude-haiku-4-5-20251001 | 2/5 | No deployed product reliably performs the full supervisory communication task; AI tools can assist with document summarization and draft communication, but actual informed briefing of personnel on implications requires a human supervisor. Products exist for policy tracking and content generation, but none substitute for the supervisor's role. |
| Technical feasibility today | claude-sonnet-5 | 3/5 | Deployed tools (legal update summarization, internal knowledge bases, LLM-based briefing generators) exist and are used in some agencies, but reliable, department-specific rollout of legal/policy communication is uneven and not yet standard practice. |
Prepare news releases and respond to police correspondence.
36CI 25–48 · exposure 38 · augmentation 75 · importance 3.5/5 · click for rater detail
Prepare news releases and respond to police correspondence.
36| Dimension | Model | Rating | Rationale |
|---|---|---|---|
| Sector adoption velocity | claude-haiku-4-5-20251001 | 2/5 | Law enforcement remains a traditionally hierarchical, risk-averse sector with limited digitization in supervisory functions. While some departments pilot communication tools, production adoption of AI-generated police correspondence is still rare; organizational inertia and liability concerns slow velocity. |
| Sector adoption velocity | claude-sonnet-5 | 2/5 | Law enforcement agencies are generally slow, under-resourced, and cautious in adopting AI tools for public-facing communications compared to fast-moving private sectors. |
| Augmentation potential | claude-haiku-4-5-20251001 | 4/5 | AI writing assistants can meaningfully accelerate news release drafting and correspondence composition by generating first drafts, organizing facts, and suggesting language, allowing supervisors to focus on legal review and institutional voice rather than writing from scratch. This represents substantial augmentation while keeping human judgment and accountability central. |
| Augmentation potential | claude-sonnet-5 | 4/5 | AI can significantly speed up drafting of press releases and correspondence responses, letting supervisors edit and approve rather than write from scratch. |
| Task automatability | claude-haiku-4-5-20251001 | 2/5 | AI can draft routine news releases and generate template responses to standard correspondence, but police communications require careful accuracy, legal compliance, and institutional voice that demand substantial human oversight. The task cannot achieve the 50% time savings at equal quality threshold without significant human revision and judgment. |
| Task automatability | claude-sonnet-5 | 3/5 | Drafting news releases and routine correspondence responses is well within current LLM capability, but requires human review for accuracy, sensitivity, and legal implications, limiting full automation of the task as a whole. |
| Adoption barriers | claude-haiku-4-5-20251001 | 4/5 | Police communications carry legal, liability, and public-relations risks that create strong incentives for human accountability. Departments typically require supervisory sign-off and may face reputational or legal exposure from factual errors or tone misstatements in official releases and responses, creating organizational friction against full automation. |
| Adoption barriers | claude-sonnet-5 | 3/5 | No licensing requirement for drafting text, but public accountability, legal liability for misstatements, and department policy on official communications create meaningful oversight friction. |
| Cost vs. human wage | claude-haiku-4-5-20251001 | 2/5 | AI inference for drafting is cheap, but integration, fact-checking, legal review, and human refinement still consume significant labor. The all-in cost of reliable police communications remains comparable to or exceeds the cost of a supervisor writing directly, especially accounting for liability and error correction. |
| Cost vs. human wage | claude-sonnet-5 | 3/5 | AI drafting is cheap per item, but the supervisor still must review, verify facts, and approve releases, so overall cost savings are moderate rather than transformative. |
| Technical feasibility today | claude-haiku-4-5-20251001 | 3/5 | AI writing tools and language models exist and can generate news release drafts and correspondence responses in production, but they require material human fact-checking, legal review, and tone adjustment for police contexts. Deployed products handle the mechanics but not the full quality and liability standards police departments require. |
| Technical feasibility today | claude-sonnet-5 | 2/5 | Generic drafting tools exist and are used ad hoc by public agencies, but no widely deployed product specifically handles police correspondence/news releases reliably in production with the accountability needed. |
Prepare budgets and manage expenditures of department funds.
36CI 25–48 · exposure 38 · augmentation 75 · importance 3.1/5 · click for rater detail
Prepare budgets and manage expenditures of department funds.
36| Dimension | Model | Rating | Rationale |
|---|---|---|---|
| Sector adoption velocity | claude-haiku-4-5-20251001 | 2/5 | Public-sector budgeting has moderate digitization, but adoption of autonomous financial decision-making is slow due to regulatory oversight, risk aversion, and established procurement processes that favor human sign-off. |
| Sector adoption velocity | claude-sonnet-5 | 2/5 | Government and law enforcement agencies are typically slow adopters of AI for financial management due to bureaucratic processes, procurement cycles, and limited digitization compared to private-sector finance. |
| Augmentation potential | claude-haiku-4-5-20251001 | 4/5 | AI can effectively assist supervisors by automating expense categorization, forecasting spending trends, generating budget reports, and flagging anomalies, meaningfully raising their productivity while they retain decision authority over allocation and strategy. |
| Augmentation potential | claude-sonnet-5 | 4/5 | AI can meaningfully assist supervisors by automating data aggregation, trend analysis, and draft budget preparation, allowing them to focus on strategic and oversight decisions. |
| Task automatability | claude-haiku-4-5-20251001 | 2/5 | Budget preparation involves routine data entry and calculation that AI could partially automate, but meaningful portions require human judgment on operational priorities, resource allocation, and departmental strategy that current systems cannot autonomously determine. |
| Task automatability | claude-sonnet-5 | 3/5 | AI can draft budget documents, forecast expenditures, and analyze spending data with significant time savings, but final judgment calls on resource allocation, political considerations, and approval remain human-driven.dd |
| Adoption barriers | claude-haiku-4-5-20251001 | 4/5 | Budget approval and financial stewardship typically require sign-off by authorized supervisory and administrative personnel under government financial controls; legal and audit requirements mandate human accountability for fund allocation decisions. |
| Adoption barriers | claude-sonnet-5 | 3/5 | Public sector budgeting requires accountability, transparency, and often sign-off by an authorized officer or elected body, creating moderate procedural and legal friction against full automation. |
| Cost vs. human wage | claude-haiku-4-5-20251001 | 2/5 | Off-the-shelf financial software and AI tools are relatively inexpensive, but the supervisor's expertise in police operations and budget strategy is highly specialized; the all-in cost of AI-plus-oversight remains comparable to or exceeds direct human labor. |
| Cost vs. human wage | claude-sonnet-5 | 3/5 | AI tools can cut time spent on data compilation and forecasting, but a supervisor or finance officer still must interpret, justify, and defend the budget, keeping overall costs roughly comparable to traditional processes. |
| Technical feasibility today | claude-haiku-4-5-20251001 | 2/5 | Financial software can assist with expense tracking and report generation, but fully autonomous budget preparation and fund management requires understanding of organizational constraints, stakeholder input, and policy compliance that no deployed product reliably handles without human oversight. |
| Technical feasibility today | claude-sonnet-5 | 3/5 | Financial planning and budgeting software with AI-assisted forecasting exists and is used in government finance offices, but police-specific budget management still relies heavily on manual review and human decision-making. |
Review contents of written orders to ensure adherence to legal requirements.
25CI 25–25 · exposure 25 · augmentation 63 · importance 4.0/5 · click for rater detail
Review contents of written orders to ensure adherence to legal requirements.
25| Dimension | Model | Rating | Rationale |
|---|---|---|---|
| Sector adoption velocity | claude-haiku-4-5-20251001 | 2/5 | Law enforcement agencies have adopted digital systems slowly and remain conservative on automation in legal and constitutional domains. Current practice relies on human chain-of-command review rather than algorithmic flagging, with limited production deployment of AI for legal compliance tasks. |
| Sector adoption velocity | claude-sonnet-5 | 2/5 | Law enforcement is a traditionally slow-adopting sector for AI in legal/compliance functions, with pilots emerging but production deployment for order review still rare. |
| Augmentation potential | claude-haiku-4-5-20251001 | 3/5 | AI could usefully assist supervisors by flagging language patterns, cross-referencing recent case law, and highlighting potential statutory conflicts, allowing the human supervisor to focus their legal judgment on higher-risk provisions. This is a credible augmentation scenario despite low automatability. |
| Augmentation potential | claude-sonnet-5 | 4/5 | AI can meaningfully assist by scanning documents for known legal red flags, inconsistencies, or missing required language, helping supervisors focus their review, even though final sign-off remains human. |
| Task automatability | claude-haiku-4-5-20251001 | 2/5 | While AI can extract and flag potentially non-compliant language in written orders, the task fundamentally requires legal expertise and judgment about nuanced statutory and constitutional requirements that vary by jurisdiction. AI assistance here is partial and would still require substantial human legal review, falling short of the 50% time-saving threshold. |
| Task automatability | claude-sonnet-5 | 2/5 | AI can flag compliance issues in written orders via document review, but final legal adherence judgment for police directives requires contextual, jurisdiction-specific expertise that current systems cannot fully replicate end-to-end.6d Time savings likely below 50% given need for careful human verification of legal nuance. |
| Adoption barriers | claude-haiku-4-5-20251001 | 4/5 | Law enforcement operates under strict regulatory frameworks, civil rights liability standards, and chain-of-command accountability requirements. A human first-line supervisor carries legal responsibility for order legality, and automation faces organizational, liability, and potentially regulatory resistance. |
| Adoption barriers | claude-sonnet-5 | 4/5 | Legal compliance sign-off for police orders typically requires an authorized supervisor or legal counsel; liability and regulatory frameworks around law enforcement conduct create strong barriers to full automation. |
| Cost vs. human wage | claude-haiku-4-5-20251001 | 2/5 | While AI inference costs are low, effective deployment would require legal oversight and integration with existing order management systems, plus liability exposure for missed non-compliance. Total cost remains comparable to or higher than a qualified legal or supervisory review. |
| Cost vs. human wage | claude-sonnet-5 | 2/5 | AI could assist cheaply in flagging language issues, but the oversight and legal liability review still requires a trained supervisor, keeping overall cost comparable to human-only review rather than dramatically cheaper. |
| Technical feasibility today | claude-haiku-4-5-20251001 | 2/5 | AI systems can highlight legal keywords and patterns, but no production system reliably reviews police orders for constitutional and statutory compliance with the error margins acceptable in law enforcement contexts. Deployed products lack the jurisdiction-specific legal knowledge and accountability this task demands. |
| Technical feasibility today | claude-sonnet-5 | 2/5 | Legal-document review AI tools exist in adjacent fields (contract review, compliance checking) but no deployed product is demonstrated specifically reviewing police written orders for legal compliance at scale in production. |
Develop, implement, and revise departmental policies and procedures.
25CI 25–25 · exposure 25 · augmentation 63 · importance 3.6/5 · click for rater detail
Develop, implement, and revise departmental policies and procedures.
25| Dimension | Model | Rating | Rationale |
|---|---|---|---|
| Sector adoption velocity | claude-haiku-4-5-20251001 | 2/5 | Law enforcement adoption of AI for policy development is minimal; agencies remain conservative about automation in governance and liability-sensitive domains. Most policy revision is still handled through traditional committee and leadership processes, with no measurable trend toward AI-driven automation in this space. |
| Sector adoption velocity | claude-sonnet-5 | 2/5 | Law enforcement agencies are historically slow and cautious adopters of AI for governance and policy functions, with adoption concentrated in narrow analytics tools rather than policymaking. |
| Augmentation potential | claude-haiku-4-5-20251001 | 3/5 | AI can usefully assist supervisors by generating policy drafts, flagging inconsistencies with existing procedures, and identifying best-practice language from comparable agencies, raising productivity on the research and drafting phases while the supervisor retains final judgment and accountability. |
| Augmentation potential | claude-sonnet-5 | 4/5 | AI can meaningfully help by drafting initial policy language, comparing against best practices or other departments' policies, and checking for internal consistency, substantially speeding up the drafting phase. |
| Task automatability | claude-haiku-4-5-20251001 | 2/5 | AI can assist in drafting policy language and identifying procedural gaps through analysis of existing documents, but the task fundamentally requires human judgment on organizational values, legal compliance, stakeholder input, and change management. End-to-end automation with 50% time savings at equal quality is infeasible because policy development is inherently deliberative and accountable. |
| Task automatability | claude-sonnet-5 | 2/5 | AI can draft policy language and summarize best practices, but the task requires judgment about local legal context, department culture, union negotiations, and accountability that current systems cannot autonomously resolve end-to-end.' |
| Adoption barriers | claude-haiku-4-5-20251001 | 4/5 | Departmental policies and procedures often require sign-off by legal counsel, civil service commissions, or senior leadership, creating formal authorization barriers. Additionally, policy effectiveness depends on organizational buy-in and legitimacy, which typically require human deliberation and accountability that automated systems cannot substitute. |
| Adoption barriers | claude-sonnet-5 | 4/5 | Police policy is subject to legal liability, union contracts, oversight boards, and public accountability requirements, meaning a sworn supervisor or command staff must formally approve and take responsibility for any policy change. |
| Cost vs. human wage | claude-haiku-4-5-20251001 | 2/5 | AI writing and analysis tools reduce drafting time, but the task still requires senior supervisory labor for review, legal vetting, stakeholder consultation, and final approval. All-in AI costs (tool subscription plus significant human oversight) are likely comparable to or exceed the marginal cost of having a supervisor draft policy iteratively. |
| Cost vs. human wage | claude-sonnet-5 | 2/5 | AI can cheaply generate draft text, but the substantial human time needed for legal review, stakeholder consultation, and approval keeps overall cost comparable to current processes rather than dramatically cheaper. |
| Technical feasibility today | claude-haiku-4-5-20251001 | 2/5 | No deployed products reliably perform departmental policy development and revision independently. AI writing tools exist but require extensive human oversight, legal review, and institutional validation; they are assistive rather than autonomous. Pilot applications exist but lack production-scale evidence. |
| Technical feasibility today | claude-sonnet-5 | 2/5 | Generic AI writing tools are used informally for drafting text, but no deployed product manages the full policy development, stakeholder review, and revision cycle for police departments reliably in production. |
Inspect facilities, supplies, vehicles, and equipment to ensure conformance to standards.
24CI 23–25 · exposure 25 · augmentation 50 · importance 3.5/5 · click for rater detail
Inspect facilities, supplies, vehicles, and equipment to ensure conformance to standards.
24| Dimension | Model | Rating | Rationale |
|---|---|---|---|
| Sector adoption velocity | claude-haiku-4-5-20251001 | 2/5 | Law enforcement agencies are slow to adopt AI into operational workflows due to liability concerns, union resistance, and limited digitization of facility management. Adoption remains mostly in pilot or planning stages rather than widespread production deployment. |
| Sector adoption velocity | claude-sonnet-5 | 1/5 | Law enforcement is a slow-adopting, highly physical and regulated sector with minimal AI deployment for facility/equipment inspection tasks. |
| Augmentation potential | claude-haiku-4-5-20251001 | 3/5 | AI-powered visual documentation and defect flagging could assist supervisors by automating photo capture and highlighting anomalies for review, but the supervisor's judgment and accountability in certifying conformance must remain central to the task. |
| Augmentation potential | claude-sonnet-5 | 3/5 | Checklists, IoT sensors, and mobile inspection apps can assist supervisors in tracking and documenting compliance, improving efficiency without replacing the human inspector. |
| Task automatability | claude-haiku-4-5-20251001 | 2/5 | Visual inspection of physical facilities, vehicles, and equipment requires navigating real-world environments and making nuanced judgments about conformance to standards. While AI vision could automate photo documentation, the task demands in-situ assessment and contextual reasoning that current systems cannot reliably perform end-to-end without substantial human oversight. |
| Task automatability | claude-sonnet-5 | 2/5 | Physical inspection of facilities, vehicles, and equipment requires on-site presence, visual/tactile assessment, and judgment calls that current AI cannot fully perform end-to-end.' |
| Adoption barriers | claude-haiku-4-5-20251001 | 4/5 | Police operational safety and equipment readiness carry regulatory oversight and liability; a first-line supervisor's sign-off on facility and vehicle compliance is often mandated by agency standards and law enforcement regulations, creating a hard barrier to full automation. |
| Adoption barriers | claude-sonnet-5 | 4/5 | Police supervisory duties involve accountability, chain-of-command sign-off, and regulatory compliance standards that typically require a certified officer to inspect and attest to conformance. |
| Cost vs. human wage | claude-haiku-4-5-20251001 | 2/5 | Supervisory inspection involves liability and stakes high enough that human sign-off remains legally and organizationally required. The cost of integrating and validating AI vision systems, plus mandatory human review, would likely exceed the loaded wage of a supervisor performing spot checks. |
| Cost vs. human wage | claude-sonnet-5 | 2/5 | Sensor/camera-based inspection systems require significant capital investment and human oversight, making costs comparable to or higher than a supervisor's routine walkthrough for this narrow task. |
| Technical feasibility today | claude-haiku-4-5-20251001 | 2/5 | Existing computer vision systems can detect some equipment defects or obvious violations from images, but deployed products lack the spatial reasoning, safety awareness, and standard-conformance knowledge needed to reliably inspect complex police facilities and vehicles in production environments. |
| Technical feasibility today | claude-sonnet-5 | 2/5 | Some computer vision tools exist for vehicle damage or equipment defect detection, but no deployed product performs comprehensive law-enforcement facility/equipment compliance inspections autonomously. |
Train staff in proper police work procedures.
21CI 16–25 · exposure 17 · augmentation 63 · importance 4.1/5 · click for rater detail
Train staff in proper police work procedures.
21| Dimension | Model | Rating | Rationale |
|---|---|---|---|
| Sector adoption velocity | claude-haiku-4-5-20251001 | 2/5 | Police departments are traditionally hierarchical, conservative, and slow to digitize; while some e-learning pilots exist, most training remains in-person and supervisor-led, with adoption lagging broader sectors. |
| Sector adoption velocity | claude-sonnet-5 | 2/5 | Law enforcement agencies are traditionally slow to adopt new technology for core operational training, with pilots for AI-assisted training tools emerging but production-scale replacement uncommon. |
| Augmentation potential | claude-haiku-4-5-20251001 | 3/5 | AI tools can assist supervisors by generating scenario-based training materials, drafting procedure guides, or creating simulations, meaningfully supporting their instructional role while the supervisor retains control and judgment. |
| Augmentation potential | claude-sonnet-5 | 4/5 | AI can meaningfully augment training via generating scenario-based content, quizzes, VR simulations, and personalized study materials, improving efficiency while supervisors remain central to delivery. |
| Task automatability | claude-haiku-4-5-20251001 | 1/5 | Training staff in police procedures requires judgment calls on individual officer needs, personalized instruction, real-time Q&A, and assessment of understanding—all deeply interactive and context-dependent work that AI cannot perform end-to-end today. |
| Task automatability | claude-sonnet-5 | 2/5 | AI can generate training materials and simulate scenarios, but delivering hands-on training, modeling procedures, and assessing trainee performance in real situations requires human judgment and presence that current AI cannot replace end-to-end.5In practice this caps time savings well below the 50% threshold for the full task. |
| Adoption barriers | claude-haiku-4-5-20251001 | 4/5 | Police training is often mandated by law and departmental policy, with legal and liability requirements that a certified supervisor must oversee and sign off on; regulatory and accountability requirements create hard adoption barriers. |
| Adoption barriers | claude-sonnet-5 | 4/5 | Police training often involves certification requirements, legal liability, use-of-force standards, and departmental accountability that mandate qualified human trainers and sign-off, creating strong institutional barriers. |
| Cost vs. human wage | claude-haiku-4-5-20251001 | 2/5 | Training is highly labor-intensive and requires experienced supervisors; while AI can draft content or deliver e-learning, the supervision and feedback components are irreplaceable, making total cost comparable or higher than human delivery. |
| Cost vs. human wage | claude-sonnet-5 | 2/5 | AI can cheaply produce training content, but the supervisory, hands-on, and certification components still require paid human trainers, so overall cost savings are limited. |
| Technical feasibility today | claude-haiku-4-5-20251001 | 2/5 | AI can generate training materials and simulations, but no deployed system reliably trains police officers on procedure with the human judgment, credibility, and accountability that the role demands; deployed products remain narrow and supplementary. |
| Technical feasibility today | claude-sonnet-5 | 2/5 | E-learning modules and AI-generated content exist and are used in some training curricula, but no deployed product independently trains police staff in procedures at scale without heavy human instructor involvement. |
Direct collection, preparation, and handling of evidence and personal property of prisoners.
6CI 0–11 · exposure 5 · augmentation 38 · importance 4.1/5 · click for rater detail
Direct collection, preparation, and handling of evidence and personal property of prisoners.
6| Dimension | Model | Rating | Rationale |
|---|---|---|---|
| Sector adoption velocity | claude-haiku-4-5-20251001 | 2/5 | Law enforcement is moderately digitized but remains resistant to automation in safety-critical, legally regulated tasks. Adoption is slow despite digitalization pressures due to liability and regulatory constraints. |
| Sector adoption velocity | claude-sonnet-5 | 1/5 | Law enforcement evidence handling is a low-digitization, physically grounded, highly regulated domain with minimal AI adoption for direct operational control. |
| Augmentation potential | claude-haiku-4-5-20251001 | 3/5 | AI can assist with evidence logging, tracking inventory, flagging missing documentation, and generating reports, meaningfully reducing administrative burden while the supervisor retains full custodial control. |
| Augmentation potential | claude-sonnet-5 | 2/5 | AI can help with digital evidence logging, inventory databases, or documentation templates, but offers little assistance to the physical directing and custody aspects of the task. |
| Task automatability | claude-haiku-4-5-20251001 | 1/5 | Evidence and property handling requires custody chain documentation, legal compliance, and human judgment on contested items. Current AI cannot physically collect/prepare materials or make legal determinations about evidence admissibility, making full automation infeasible. |
| Task automatability | claude-sonnet-5 | 1/5 | This requires physical custody, chain-of-custody verification, and direct supervisory judgment over evidence handling and prisoner property, none of which current AI can perform end-to-end.atab |
| Adoption barriers | claude-haiku-4-5-20251001 | 5/5 | Evidence handling is governed by strict chain-of-custody legal requirements and rules of evidence; a licensed human supervisor must directly oversee and assume responsibility for evidence integrity, making automation legally prohibited. |
| Adoption barriers | claude-sonnet-5 | 5/5 | Chain-of-custody law, evidentiary admissibility rules, and departmental accountability structures require a sworn, authorized officer to direct and be responsible for evidence handling. |
| Cost vs. human wage | claude-haiku-4-5-20251001 | 2/5 | AI documentation assistance reduces some clerical work, but the core task—secure physical handling and custody chain responsibility—cannot be delegated to systems, limiting cost advantage to partial augmentation only. |
| Cost vs. human wage | claude-sonnet-5 | 1/5 | There is no viable AI substitute performing this physical directive task, so cost comparison favors the human by default. |
| Technical feasibility today | claude-haiku-4-5-20251001 | 2/5 | AI can assist with documentation and database entry, but no deployed system reliably handles the full task of collecting, preparing, and managing evidence chains with the legal accountability required in law enforcement. |
| Technical feasibility today | claude-sonnet-5 | 1/5 | No deployed AI product directs physical evidence collection or prisoner property handling; this remains a purely human, on-scene supervisory function. |
Explain police operations to subordinates to assist them in performing their job duties.
5CI 0–10 · exposure 5 · augmentation 38 · importance 4.1/5 · click for rater detail
Explain police operations to subordinates to assist them in performing their job duties.
5| Dimension | Model | Rating | Rationale |
|---|---|---|---|
| Sector adoption velocity | claude-haiku-4-5-20251001 | 1/5 | Police departments are traditionally hierarchical and risk-averse; supervisory communication of operations is a core human leadership function with high liability stakes, and adoption of AI for this role is negligible in practice. |
| Sector adoption velocity | claude-sonnet-5 | 1/5 | Law enforcement is a low-digitization, high physical-presence sector with slow AI adoption in core supervisory and operational command functions. |
| Augmentation potential | claude-haiku-4-5-20251001 | 2/5 | AI could assist a supervisor by drafting explanations of standard procedures or policies, but the supervisory task inherently involves judgment about what to communicate and responsive dialogue, limiting meaningful augmentation to narrow document-preparation aspects. |
| Augmentation potential | claude-sonnet-5 | 3/5 | AI tools (e.g., knowledge bases, training materials, policy summaries) can help supervisors prepare explanations and training content, improving efficiency without replacing the interpersonal supervisory role. |
| Task automatability | claude-haiku-4-5-20251001 | 1/5 | Explaining police operations to subordinates requires real-time responsiveness to individual questions, contextual judgment about what information each officer needs, and adaptive communication based on their experience level and current situation. Current AI cannot reliably perform this supervisory communication function end-to-end. |
| Task automatability | claude-sonnet-5 | 1/5 | This requires live, context-specific supervisory communication, judgment, and authority relationships that current AI cannot replicate end-to-end; at most AI could help draft reference materials.value cannot substitute for the supervisory act itself.' |
| Adoption barriers | claude-haiku-4-5-20251001 | 5/5 | Law enforcement has strict chain-of-command and accountability structures; a licensed, responsible supervisor must legally perform operational briefings and explanations to ensure reliability, liability coverage, and disciplinary authority. Regulatory and organizational requirements make substitution infeasible. |
| Adoption barriers | claude-sonnet-5 | 4/5 | Police supervision involves chain-of-command authority, accountability, and legal/organizational structures requiring a sworn, authorized human supervisor to direct subordinates. |
| Cost vs. human wage | claude-haiku-4-5-20251001 | 1/5 | The cost of AI systems to generate and manage contextually appropriate explanations, plus the oversight needed to ensure accuracy and appropriateness in a law-enforcement setting, exceeds the cost of a supervisor performing this core leadership function. |
| Cost vs. human wage | claude-sonnet-5 | 1/5 | There is no AI substitute performing this supervisory task, so no meaningful cost comparison exists; the human supervisor remains the only viable option. |
| Technical feasibility today | claude-haiku-4-5-20251001 | 2/5 | While AI can generate generic explanations of police procedures or policies from documents, no deployed product reliably performs the interactive, supervisory role of explaining operations to specific subordinates in context. Some tools might assist with drafting explanations but do not replace the supervisor doing this task. |
| Technical feasibility today | claude-sonnet-5 | 1/5 | No deployed product performs live supervisory explanation of police operations to subordinates; this remains a human interpersonal management function. |
Monitor and evaluate the job performance of subordinates, and authorize promotions and transfers.
4CI 0–7 · exposure 5 · augmentation 38 · importance 4.0/5 · click for rater detail
Monitor and evaluate the job performance of subordinates, and authorize promotions and transfers.
4| Dimension | Model | Rating | Rationale |
|---|---|---|---|
| Sector adoption velocity | claude-haiku-4-5-20251001 | 2/5 | Law enforcement remains a traditionally hierarchical, human-centered sector with strong union involvement and regulatory scrutiny of personnel decisions. Adoption of AI-driven evaluation or authorization is slow; most agencies still rely on manual performance reviews and supervisor judgment. |
| Sector adoption velocity | claude-sonnet-5 | 1/5 | Policing is a low-digitization, high human-contact sector with minimal AI adoption in personnel management and promotion decisions. |
| Augmentation potential | claude-haiku-4-5-20251001 | 3/5 | AI can assist by surfacing performance data (attendance, complaints, training completion) and highlighting patterns, helping supervisors make more informed decisions. However, the core evaluative and authorization acts remain human-centered, limiting transformative augmentation. |
| Augmentation potential | claude-sonnet-5 | 2/5 | AI can help compile performance metrics, incident reports, or attendance data to inform evaluations, but the judgment and authorization remain human-driven with limited AI contribution. |
| Task automatability | claude-haiku-4-5-20251001 | 1/5 | Monitoring and evaluating subordinate performance, and authorizing promotions and transfers, requires sustained human judgment about individual capabilities, organizational fit, and fairness—decisions that carry legal and career consequences. Current AI cannot reliably assess nuanced performance or make defensible personnel decisions end-to-end. |
| Task automatability | claude-sonnet-5 | 1/5 | Evaluating officer performance and authorizing promotions requires nuanced judgment about conduct, leadership potential, and departmental fit that current AI cannot reliably assess or decide end-to-end. |
| Adoption barriers | claude-haiku-4-5-20251001 | 5/5 | Employment law (Equal Employment Opportunity, union contracts, civil service rules) typically requires a human supervisor to document and authorize personnel actions. Liability for wrongful promotion or transfer rests with the employing agency, creating a hard legal barrier to full automation. |
| Adoption barriers | claude-sonnet-5 | 5/5 | Promotions and personnel decisions in law enforcement are governed by civil service rules, union contracts, and command authority requiring a human supervisor's formal sign-off and accountability. |
| Cost vs. human wage | claude-haiku-4-5-20251001 | 1/5 | AI systems for HR analytics or performance tracking add cost (licensing, integration, oversight) without displacing the supervisor role, which remains legally and ethically necessary. The supervisor wage is the marginal cost; AI overhead makes automation uneconomical. |
| Cost vs. human wage | claude-sonnet-5 | 1/5 | There is no viable AI substitute performing this function, so cost comparison favors the human supervisor by default. |
| Technical feasibility today | claude-haiku-4-5-20251001 | 2/5 | While AI can assist with data aggregation (e.g., flagging attendance records or citation metrics), no deployed system makes promotion or transfer authorization decisions independently; these remain human-driven in law enforcement. Some tools exist for performance tracking but none replace the supervisor's evaluative function. |
| Technical feasibility today | claude-sonnet-5 | 1/5 | No deployed product performs supervisory performance evaluation and promotion authorization for police personnel; this remains a human managerial function. |
Supervise and coordinate the investigation of criminal cases, offering guidance and expertise to investigators, and ensuring that procedures are conducted in accordance with laws and regulations.
0CI 0–0 · exposure 0 · augmentation 38 · importance 4.3/5 · click for rater detail
Supervise and coordinate the investigation of criminal cases, offering guidance and expertise to investigators, and ensuring that procedures are conducted in accordance with laws and regulations.
0| Dimension | Model | Rating | Rationale |
|---|---|---|---|
| Sector adoption velocity | claude-haiku-4-5-20251001 | 1/5 | Law enforcement adoption of AI for supervisory and case-oversight functions remains minimal in practice. Pilot programs exist, but production deployment of AI supervisors is negligible; organizational culture, liability concerns, and regulatory conservatism in policing create structural lag. |
| Sector adoption velocity | claude-sonnet-5 | 1/5 | Law enforcement supervisory functions are a low-digitization, highly regulated, human-contact-intensive sector with minimal AI adoption in command roles. |
| Augmentation potential | claude-haiku-4-5-20251001 | 2/5 | AI can assist supervisors with administrative tasks (case scheduling, file organization, flagging potential legal issues) or suggesting investigative leads, but it offers limited augmentation to the core supervisory, mentoring, and case-judgment functions that define this role. |
| Augmentation potential | claude-sonnet-5 | 3/5 | AI tools can help summarize case files, flag compliance issues, or search records to support the supervisor's oversight, though the core coordination and judgment remain human-driven. |
| Task automatability | claude-haiku-4-5-20251001 | 1/5 | This task requires real-time human judgment about investigative strategy, legal compliance nuance, and personnel management that cannot be reliably automated end-to-end. Current AI cannot supervise humans, make case-level decisions, or ensure investigative procedures meet jurisdiction-specific legal standards without continuous human oversight. |
| Task automatability | claude-sonnet-5 | 1/5 | This requires real-time human judgment, legal accountability, personnel management, and situational leadership over sworn officers that current AI cannot perform end-to-end. |
| Adoption barriers | claude-haiku-4-5-20251001 | 5/5 | Strong legal and regulatory barriers protect this task: supervisors have statutory authority, accountability for case integrity, and must sign off on investigative legality. Liability and chain-of-custody requirements, combined with organizational and union structures, create hard barriers to substitution. |
| Adoption barriers | claude-sonnet-5 | 5/5 | Command authority over investigations is legally vested in sworn, certified law enforcement supervisors, with strict chain-of-command, evidentiary, and due-process requirements barring automation. |
| Cost vs. human wage | claude-haiku-4-5-20251001 | 1/5 | The cost of an AI system to replace a police supervisor (infrastructure, integration, legal review, continuous monitoring for bias/compliance) would far exceed the salary of a first-line supervisor, especially given liability exposure in law enforcement. |
| Cost vs. human wage | claude-sonnet-5 | 1/5 | There is no viable AI substitute for the supervisory role, so cost comparison favors the human by default since the AI alternative does not exist for the core task. |
| Technical feasibility today | claude-haiku-4-5-20251001 | 1/5 | No deployed product reliably performs supervisory coordination of criminal investigations or offers investigative guidance in production systems. While AI can assist with evidence organization or pattern detection, the supervisory and compliance-oversight functions remain research-stage. |
| Technical feasibility today | claude-sonnet-5 | 1/5 | No deployed product supervises criminal investigations or directs investigators; AI is at most a research-stage decision-support concept in this domain. |
Investigate and resolve personnel problems within organization and charges of misconduct against staff.
0CI 0–0 · exposure 0 · augmentation 25 · importance 4.1/5 · click for rater detail
Investigate and resolve personnel problems within organization and charges of misconduct against staff.
0| Dimension | Model | Rating | Rationale |
|---|---|---|---|
| Sector adoption velocity | claude-haiku-4-5-20251001 | 1/5 | Law enforcement and public sector organizations are among the slowest to adopt automation in sensitive personnel functions; misconduct investigations carry high legal and reputational stakes, making adoption of AI-driven systems negligible. |
| Sector adoption velocity | claude-sonnet-5 | 1/5 | Law enforcement HR and internal affairs functions are slow-adopting, highly procedural, and resistant to AI-driven decision-making due to legal liability and union oversight. |
| Augmentation potential | claude-haiku-4-5-20251001 | 2/5 | AI could assist with document organization, timeline construction, or flagging interview patterns, but the core investigative task—interviewing, judgment, and legal accountability—remains fundamentally human; augmentation is marginal. |
| Augmentation potential | claude-sonnet-5 | 2/5 | AI can help organize case files, transcribe interviews, or search policy documents, but offers only marginal assistance to the core judgment-based investigative and resolution work. |
| Task automatability | claude-haiku-4-5-20251001 | 1/5 | Investigating and resolving personnel problems and misconduct charges requires evaluating complex human behavior, interpersonal dynamics, legal implications, and organizational context that demands discretionary judgment and accountability. Current AI systems cannot reliably conduct investigations requiring witness interviews, credibility assessment, and legally defensible fact-finding. |
| Task automatability | claude-sonnet-5 | 1/5 | Investigating personnel misconduct requires interviewing witnesses, judging credibility, applying disciplinary policy, and making authoritative decisions with legal and career consequences—none of which current AI can perform end-to-end. |
| Adoption barriers | claude-haiku-4-5-20251001 | 5/5 | Strong legal and regulatory barriers exist: employment law requires documented investigation processes, many jurisdictions mandate specific human oversight, and liability for wrongful discipline falls on the organization and its authorized agents, creating a hard requirement for human responsibility. |
| Adoption barriers | claude-sonnet-5 | 5/5 | Police discipline processes are governed by union contracts, civil service law, due process requirements, and chain-of-command authority, requiring a certified human supervisor to conduct and sign off on investigations. |
| Cost vs. human wage | claude-haiku-4-5-20251001 | 1/5 | AI oversight and integration costs for misconduct investigations would be high due to required human oversight, legal review, and documentation; the task cannot achieve cost efficiency gains over human investigators who are legally accountable for findings. |
| Cost vs. human wage | claude-sonnet-5 | 1/5 | There is no viable AI substitute performing this task, so cost comparison favors the human by default; any AI use is limited to note-taking or research support. |
| Technical feasibility today | claude-haiku-4-5-20251001 | 1/5 | No deployed product reliably performs personnel misconduct investigations end-to-end; this remains the domain of trained human investigators and HR professionals because of liability exposure, legal privilege requirements, and the need for human judgment in sensitive employment matters. |
| Technical feasibility today | claude-sonnet-5 | 1/5 | No deployed product conducts internal affairs or personnel misconduct investigations autonomously; this remains firmly a human supervisory function. |
Discipline staff for violation of department rules and regulations.
0CI 0–0 · exposure 0 · augmentation 25 · importance 4.1/5 · click for rater detail
Discipline staff for violation of department rules and regulations.
0| Dimension | Model | Rating | Rationale |
|---|---|---|---|
| Sector adoption velocity | claude-haiku-4-5-20251001 | 1/5 | Public-sector law enforcement agencies are highly conservative, heavily unionized, and subject to strict civil-service and legal frameworks that strongly protect human decision-making in personnel matters. Adoption of AI for discipline would face intense resistance. |
| Sector adoption velocity | claude-sonnet-5 | 1/5 | Law enforcement is a slow-adopting, highly regulated sector with strong human-in-command norms for personnel actions, and no evidence of AI-driven disciplinary processes emerging. |
| Augmentation potential | claude-haiku-4-5-20251001 | 2/5 | AI might assist with flagging policy violations or summarizing prior discipline records, but the core task—making a judgment call with legal and human consequences—resists meaningful augmentation without introducing unacceptable liability. |
| Augmentation potential | claude-sonnet-5 | 2/5 | AI could help draft documentation, summarize incident reports, or check policy compliance, but the actual decision and delivery of discipline remain a human-only judgment task. |
| Task automatability | claude-haiku-4-5-20251001 | 1/5 | Disciplining staff requires contextual judgment, understanding of individual circumstances, organizational politics, legal exposure, and human-centered decision-making that current AI cannot perform end-to-end. The task is inherently discretionary and situational rather than procedural. |
| Task automatability | claude-sonnet-5 | 1/5 | Disciplining staff requires investigating conduct, exercising judgment about intent and context, and delivering interpersonal consequences—none of which current AI can execute end-to-end. |
| Adoption barriers | claude-haiku-4-5-20251001 | 5/5 | Disciplinary decisions carry legal liability for wrongful termination, discrimination, and due-process violations; human supervisors are legally and organizationally accountable for these decisions. Regulatory frameworks, union agreements, and civil-service protections explicitly require human judgment and accountability. |
| Adoption barriers | claude-sonnet-5 | 5/5 | Discipline in policing involves union contracts, due process rights, civil service protections, and legal liability, requiring an authorized human supervisor to make and document the decision. |
| Cost vs. human wage | claude-haiku-4-5-20251001 | 1/5 | AI systems for this task would still require substantial human oversight, legal review, and final decision-making by the supervisor, making the all-in cost comparable to or higher than direct human performance without meaningful time savings. |
| Cost vs. human wage | claude-sonnet-5 | 1/5 | There is no viable AI substitute performing this task, so cost comparison favors the human by default since AI cannot deliver the output at all. |
| Technical feasibility today | claude-haiku-4-5-20251001 | 1/5 | No deployed product reliably performs disciplinary decisions for personnel management today. This task requires nuanced legal compliance, organizational precedent, and accountability that lie well outside current AI capabilities and would face severe liability if automated. |
| Technical feasibility today | claude-sonnet-5 | 1/5 | No deployed product performs employee discipline actions in law enforcement settings; this remains entirely a human managerial function. |
Conduct raids and order detention of witnesses and suspects for questioning.
0CI 0–0 · exposure 0 · augmentation 25 · importance 3.9/5 · click for rater detail
Conduct raids and order detention of witnesses and suspects for questioning.
0| Dimension | Model | Rating | Rationale |
|---|---|---|---|
| Sector adoption velocity | claude-haiku-4-5-20251001 | 1/5 | Law enforcement automation in this domain remains minimal because detention and raid decisions are legally vested in human supervisors; no production systems are displacing human decision-makers in warrant execution or suspect detention. |
| Sector adoption velocity | claude-sonnet-5 | 1/5 | Law enforcement's physical, high-stakes tactical operations show essentially no AI adoption trend for this specific task. |
| Augmentation potential | claude-haiku-4-5-20251001 | 2/5 | AI could assist with planning raids via data aggregation or suspect intelligence, but the core task of conducting raids and ordering detention cannot be materially aided by AI while the human remains in control. |
| Augmentation potential | claude-sonnet-5 | 2/5 | AI can assist with intelligence gathering, suspect identification, or planning logistics beforehand, but offers minimal direct assistance during the raid or detention decision itself. |
| Task automatability | claude-haiku-4-5-20251001 | 1/5 | Conducting raids and ordering detention of witnesses requires physical presence, split-second judgment in dynamic situations, and authority that cannot be delegated to AI systems. No current AI system can physically execute raids or make binding detention decisions. |
| Task automatability | claude-sonnet-5 | 1/5 | Conducting raids and ordering detention involves physical law enforcement action, on-the-spot tactical judgment, and legal authority that AI cannot perform or replace.' |
| Adoption barriers | claude-haiku-4-5-20251001 | 5/5 | This task is protected by strict legal and regulatory barriers: Fourth Amendment protections, state criminal procedure codes, and constitutional requirements that a qualified human officer must authorize and execute detention decisions. |
| Adoption barriers | claude-sonnet-5 | 5/5 | Legal authority to detain individuals and conduct raids is strictly limited to sworn, licensed law enforcement officers under constitutional and statutory constraints, making this a hard-barrier task. |
| Cost vs. human wage | claude-haiku-4-5-20251001 | 1/5 | The task requires human police supervisors with full legal authority and liability; AI systems cannot replace this and would add cost via oversight without reducing the need for personnel. |
| Cost vs. human wage | claude-sonnet-5 | 1/5 | There is no AI substitute performing this physical, authority-based task, so no meaningful cost comparison exists—human officers are required. |
| Technical feasibility today | claude-haiku-4-5-20251001 | 1/5 | No deployed AI product performs this task; it involves criminal justice decisions, physical enforcement, and legal authority that require human judgment and accountability in real-world law enforcement. |
| Technical feasibility today | claude-sonnet-5 | 1/5 | No deployed AI product conducts raids or issues detention orders; this remains entirely a human, in-person law enforcement function. |
Cooperate with court personnel and officials from other law enforcement agencies and testify in court, as necessary.
0CI 0–0 · exposure 0 · augmentation 38 · importance 3.9/5 · click for rater detail
Cooperate with court personnel and officials from other law enforcement agencies and testify in court, as necessary.
0| Dimension | Model | Rating | Rationale |
|---|---|---|---|
| Sector adoption velocity | claude-haiku-4-5-20251001 | 1/5 | This task occurs in heavily regulated legal and law enforcement contexts with minimal digital transformation. No measurable AI adoption is occurring because the task is legally reserved to humans. |
| Sector adoption velocity | claude-sonnet-5 | 1/5 | Law enforcement and judicial systems are slow-adopting sectors for AI in core legal processes like testimony and inter-agency liaison work. |
| Augmentation potential | claude-haiku-4-5-20251001 | 2/5 | AI might assist with case-document review or evidence organization before court, but cannot assist with the core act of testimony, cooperation with courts, or legal judgment itself. Augmentation is minimal. |
| Augmentation potential | claude-sonnet-5 | 3/5 | AI can help supervisors prepare testimony, organize case files, and draft correspondence with other agencies, but cannot replace the interpersonal and legal core of the task. |
| Task automatability | claude-haiku-4-5-20251001 | 1/5 | Testifying in court and cooperating with legal officials requires live human presence, adversarial judgment, credibility assessment, and courtroom procedure compliance—none of which can be automated or delegated to AI today. This task is fundamentally dependent on human legal authority and presence. |
| Task automatability | claude-sonnet-5 | 1/5 | Testifying in court and coordinating with officials requires a legally accountable human presence, personal credibility, and live sworn testimony that AI cannot perform end-to-end today.' |
| Adoption barriers | claude-haiku-4-5-20251001 | 5/5 | This task has hard legal barriers: only humans can testify under oath, courts require sworn officials, and law is the regulation itself. Automation is legally prohibited. |
| Adoption barriers | claude-sonnet-5 | 5/5 | Courtroom testimony legally requires a sworn human witness with personal knowledge and accountability; this is a hard legal/procedural barrier to any automation. |
| Cost vs. human wage | claude-haiku-4-5-20251001 | 1/5 | AI cannot perform this task at any cost, since human presence and sworn testimony are legally mandated. The comparison is not meaningful. |
| Cost vs. human wage | claude-sonnet-5 | 1/5 | There is no viable AI alternative to human testimony/cooperation, so cost comparison favors the human by default as AI cannot substitute at all. |
| Technical feasibility today | claude-haiku-4-5-20251001 | 1/5 | No deployed AI system can perform testimony or legal cooperation; courts require licensed attorneys or sworn personnel. This task is legally reserved to humans and not deployable by AI. |
| Technical feasibility today | claude-sonnet-5 | 1/5 | No deployed product testifies in court or serves as a legal liaison; this remains entirely a human function with no production AI substitutes. |
Meet with civic, educational, and community groups to develop community programs and events, and to discuss law enforcement subjects.
0CI 0–0 · exposure 0 · augmentation 38 · importance 3.6/5 · click for rater detail
Meet with civic, educational, and community groups to develop community programs and events, and to discuss law enforcement subjects.
0| Dimension | Model | Rating | Rationale |
|---|---|---|---|
| Sector adoption velocity | claude-haiku-4-5-20251001 | 1/5 | Police departments have not and are unlikely to adopt AI to replace or conduct community outreach meetings. This work is explicitly relational and community-facing, domains where human supervisors are strongly preferred and required for legitimacy. |
| Sector adoption velocity | claude-sonnet-5 | 1/5 | Public safety/community relations functions in government are slow-adopting sectors with minimal AI penetration into interpersonal civic engagement work. |
| Augmentation potential | claude-haiku-4-5-20251001 | 2/5 | AI could assist minimally by preparing meeting agendas, summarizing community feedback, or drafting post-meeting reports, but the core task of meeting and engaging stakeholders remains fundamentally human-centered. |
| Augmentation potential | claude-sonnet-5 | 3/5 | AI can help prepare talking points, summarize crime data/statistics for presentations, and draft event materials, but cannot replace the live interpersonal engagement itself. |
| Task automatability | claude-haiku-4-5-20251001 | 1/5 | This task requires genuine interpersonal engagement, negotiation, and relationship-building with community stakeholders to develop collaborative programs. Current AI systems cannot authentically participate in community meetings or build the trust relationships necessary for effective civic outreach. |
| Task automatability | claude-sonnet-5 | 1/5 | This requires in-person relationship building, trust, live facilitation, and responsiveness to community concerns that AI cannot perform end-to-end today. |
| Adoption barriers | claude-haiku-4-5-20251001 | 5/5 | Strong legal and organizational barriers exist: only authorized law enforcement personnel can represent the police department to the public, speak on behalf of the agency, and make commitments on community programs. This role requires human accountability and legal standing. |
| Adoption barriers | claude-sonnet-5 | 5/5 | A sworn, authorized supervisor must represent the department, build community trust, and be accountable for public statements and commitments made on behalf of law enforcement. |
| Cost vs. human wage | claude-haiku-4-5-20251001 | 1/5 | An AI system would not reduce costs for this task; a supervisor must still attend and engage directly. Integration would add overhead rather than substitute for the human's presence and decision-making. |
| Cost vs. human wage | claude-sonnet-5 | 1/5 | There is no AI substitute performing this function, so cost comparison favors the human by default since the task cannot be delegated to AI. |
| Technical feasibility today | claude-haiku-4-5-20251001 | 1/5 | No deployed AI system can independently conduct community meetings, negotiate program details, or represent law enforcement in civic dialogue. This requires human judgment, emotional intelligence, and the legal authority vested in a police supervisor. |
| Technical feasibility today | claude-sonnet-5 | 1/5 | No deployed product conducts community liaison meetings or represents a police department in live civic engagement; this is entirely human-performed. |
Direct release or transfer of prisoners.
0CI 0–0 · exposure 0 · augmentation 25 · importance 3.5/5 · click for rater detail
Direct release or transfer of prisoners.
0| Dimension | Model | Rating | Rationale |
|---|---|---|---|
| Sector adoption velocity | claude-haiku-4-5-20251001 | 1/5 | Prisoner management operates in highly regulated, conservative law-enforcement contexts with strict hierarchical control and legal requirements for human decision-making at every step. |
| Sector adoption velocity | claude-sonnet-5 | 1/5 | Law enforcement and corrections are slow-adopting, highly regulated sectors with minimal AI deployment in custodial decision-making. |
| Augmentation potential | claude-haiku-4-5-20251001 | 2/5 | AI might assist by flagging missing documentation or processing paperwork templates, but the core supervisory decision and authorization must remain with the human supervisor; augmentation potential is minimal. |
| Augmentation potential | claude-sonnet-5 | 2/5 | AI can help manage records, scheduling, and documentation supporting release/transfer decisions, but offers minimal assistance to the core directive and judgment task itself. |
| Task automatability | claude-haiku-4-5-20251001 | 1/5 | Directing prisoner release or transfer requires custody decisions and legal authority that are inherently human functions. Current AI systems cannot make binding decisions about detainment status, sign custody transfer documents, or assume legal responsibility. |
| Task automatability | claude-sonnet-5 | 1/5 | Directing prisoner release/transfer requires legal authority, judgment calls on custody status, and accountability that AI cannot exercise; no end-to-end automation is plausible today. |
| Adoption barriers | claude-haiku-4-5-20251001 | 5/5 | Strong legal and regulatory barriers exist: only authorized law enforcement personnel with proper licensing can direct prisoner custody decisions, and liability for wrongful release or improper transfer falls on the responsible human official. |
| Adoption barriers | claude-sonnet-5 | 5/5 | This is a legally authorized law-enforcement function with strict chain-of-custody, liability, and statutory requirements mandating a sworn officer/supervisor to direct release or transfer. |
| Cost vs. human wage | claude-haiku-4-5-20251001 | 1/5 | An AI system cannot reduce costs here because it cannot legally or reliably perform the task end-to-end. Human supervisors must remain in the loop, so no cost displacement occurs. |
| Cost vs. human wage | claude-sonnet-5 | 1/5 | There is no viable AI substitute performing this task, so cost comparison favors the human by default; AI cannot deliver the output at all. |
| Technical feasibility today | claude-haiku-4-5-20251001 | 1/5 | No deployed product autonomously directs prisoner release or transfer decisions. These actions require human supervisory authority, chain-of-custody compliance, and legal accountability that AI systems are not designed or permitted to perform. |
| Technical feasibility today | claude-sonnet-5 | 1/5 | No deployed product performs this supervisory/authorization function; at most software tracks records, but the directive act remains human-only. |
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.