Police Identification and Records Officers

33-3021.02
Median wage $93,790/yr114,430 employed (US)Rank #696 of 923 scored · top 75% by substitution

Collect evidence at crime scene, classify and identify fingerprints, and photograph evidence for use in criminal and civil cases.

Sub-scores

0–100 · band = confidence interval from rater disagreement

Substitution19
Exposure21
Augmentation45

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

16 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%22

panel mean rating 1.9/5 → substitution pressure 22/100

Technical feasibility todayw 20%21

panel mean rating 1.9/5 → substitution pressure 21/100

Cost vs. human wagew 15%19

panel mean rating 1.8/5 → substitution pressure 19/100

Adoption barriersw 20%inverted — strong barriers lower the score13

panel mean rating 4.5/5 (barrier strength) → substitution pressure 13/100

Sector adoption velocityw 10%17

panel mean rating 1.7/5 → substitution pressure 17/100

Task breakdown (16 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.

Identify, compare, classify, and file fingerprints, using systems such as Automated Fingerprint Identification System (AFIS) or the Henry Classification System.

66

CI 50–81 · exposure 72 · augmentation 88 · importance 4.3/5 · click for rater detail

Panel ratings per dimension for this task
DimensionModelRatingRationale
Sector adoption velocityclaude-haiku-4-5-202510015/5AFIS adoption is mature and widespread across U.S. and international law enforcement; fingerprint databases are digitized and integrated into modern policing infrastructure, with fingerprint matching already heavily automated in production systems.
Sector adoption velocityclaude-sonnet-53/5Law enforcement agencies have adopted AFIS widely over decades, but overall public-sector digitization and AI adoption in policing lags behind private-sector professional services.
Augmentation potentialclaude-haiku-4-5-202510014/5AI fingerprint systems substantially augment officer productivity by rapidly screening candidates and highlighting potential matches, allowing human officers to focus verification and decision-making on the most probable candidates rather than manual comparison.
Augmentation potentialclaude-sonnet-55/5AFIS dramatically accelerates fingerprint search and narrows candidate matches, greatly boosting examiner productivity while leaving final verification to humans.
Task automatabilityclaude-haiku-4-5-202510014/5Modern AFIS systems and AI-based fingerprint recognition can automatically classify, compare, and match fingerprints against databases with high accuracy (>99% in many cases), meeting the 50% time-saving threshold. Filing and identification of new prints can be largely automated, though some edge cases and quality control may still require human review.
Task automatabilityclaude-sonnet-53/5AFIS automation already handles matching and classification computationally, but final identification confirmation, chain-of-custody, and court-defensible verification still require trained human examiners, so full end-to-end automation at equal quality isn't yet standard.
Adoption barriersclaude-haiku-4-5-202510013/5While fingerprint identification is a routine administrative task without legal licensing requirements for the AI itself, law enforcement organizations often have procedural oversight, chain-of-custody requirements, and institutional preferences for human verification of criminal matches, creating moderate adoption friction.
Adoption barriersclaude-sonnet-54/5Forensic identification results often must be certified by qualified examiners for legal admissibility, and chain-of-custody/evidentiary standards create strong procedural and legal barriers to full automation.
Cost vs. human wageclaude-haiku-4-5-202510015/5AFIS and fingerprint AI systems operate at minimal marginal cost per task (database lookup and matching), making them orders of magnitude cheaper than paying a full officer's loaded wage for the same throughput.
Cost vs. human wageclaude-sonnet-53/5AFIS software and infrastructure require significant upfront investment and maintenance, but per-search costs are low once deployed, making it roughly comparable to or somewhat cheaper than fully manual classification over volume.
Technical feasibility todayclaude-haiku-4-5-202510015/5AFIS and commercial fingerprint recognition systems are deployed in law enforcement agencies worldwide and operate reliably at scale; these mature products perform fingerprint identification, comparison, and classification in production with established accuracy standards.
Technical feasibility todayclaude-sonnet-54/5AFIS systems are mature, widely deployed products used daily in law enforcement agencies for fingerprint search and classification, though human examiners verify matches before official use.

Take fingerprints.

52

CI 16–87 · exposure 55 · augmentation 63 · importance 4.0/5 · click for rater detail

Panel ratings per dimension for this task
DimensionModelRatingRationale
Sector adoption velocityclaude-haiku-4-5-202510014/5Fingerprint automation is already deeply embedded in law enforcement and border agencies globally; AFIS with automated capture is the operational norm rather than emerging technology. Deployment is fast across well-resourced public sectors, though some smaller or less-digitized agencies lag.
Sector adoption velocityclaude-sonnet-52/5Law enforcement and government records functions are historically slow adopters of new automation technology due to regulatory, budgetary, and procedural constraints.
Augmentation potentialclaude-haiku-4-5-202510014/5AI-assisted capture (real-time quality feedback, automatic re-scanning of poor-quality prints, guided hand positioning) significantly enhances officer productivity and print quality, keeping the human in a supervisory loop while automating the repetitive mechanical work.
Augmentation potentialclaude-sonnet-53/5Digital scanning and AI-assisted print quality checks and matching software help officers work more efficiently, though the physical fingerprinting act itself sees limited AI augmentation.
Task automatabilityclaude-haiku-4-5-202510015/5Fingerprint capture is now fully automatable: modern AFIS (Automated Fingerprint Identification Systems) with digital scanners and AI-enhanced quality checks can acquire, verify, and process fingerprints with >50% time savings and equal or higher accuracy than manual inking and rolling. The task is mechanical and deterministic, requiring only image acquisition and basic validation.
Task automatabilityclaude-sonnet-51/5Taking fingerprints requires physical manipulation of a subject's hands with specialized ink/scanner equipment, which current AI systems cannot perform as they lack physical embodiment.'
Adoption barriersclaude-haiku-4-5-202510013/5While no law explicitly mandates human fingerprint-taking, some jurisdictions and contexts (arrestee processing, sensitive populations) have procedural or liability preferences for human oversight and verification. Regulatory requirements for chain-of-custody documentation create modest friction but do not legally require human capture.
Adoption barriersclaude-sonnet-54/5Fingerprinting for law enforcement records typically requires an authorized officer for chain-of-custody, legal admissibility, and identification integrity, creating strong procedural and legal barriers to full automation.
Cost vs. human wageclaude-haiku-4-5-202510015/5Automated fingerprint capture hardware (scanners, kiosks) costs are low and amortize rapidly across thousands of individuals, making per-capture cost negligible compared to the loaded hourly wage of a trained officer (~$35–50/hour). Inference and maintenance are minimal.
Cost vs. human wageclaude-sonnet-52/5Automated live-scan fingerprint machines reduce some labor cost per capture, but a trained officer must still be present to operate equipment, verify quality, and handle the subject, keeping overall costs comparable to human labor.
Technical feasibility todayclaude-haiku-4-5-202510015/5Deployed fingerprint capture systems are production-standard in law enforcement, border control, and identity verification. Commercial systems (e.g., NEC, Gemalto, Aware) reliably capture and validate prints at scale with established error rates and integration into AFIS workflows.
Technical feasibility todayclaude-sonnet-52/5Digital fingerprint scanners with software assistance exist and are deployed, but the actual physical act of capturing prints from a live subject still requires a human operator to position hands and manage the process.

Process film and prints from crime or accident scenes.

46

CI 16–76 · exposure 53 · augmentation 50 · importance 4.1/5 · click for rater detail

Panel ratings per dimension for this task
DimensionModelRatingRationale
Sector adoption velocityclaude-haiku-4-5-202510013/5Law enforcement agencies have piloted automated fingerprint and document processing systems (e.g., AFIS variants), but deployment remains inconsistent across departments; adoption is steady but not yet deep or uniform, reflecting budget constraints and procedural conservatism.
Sector adoption velocityclaude-sonnet-52/5Law enforcement and forensic labs are typically slow adopters of AI for evidence-critical physical processes due to regulatory and reliability concerns.
Augmentation potentialclaude-haiku-4-5-202510014/5AI tools for image enhancement, rapid fingerprint matching, and automated metadata tagging substantially accelerate records officers' work while maintaining human oversight and judgment on verification and case linkage decisions.
Augmentation potentialclaude-sonnet-52/5AI can help with digital image enhancement, tagging, or database matching afterward, but offers little assistance to the core physical processing step itself.
Task automatabilityclaude-haiku-4-5-202510015/5Modern computer vision and image processing systems can automatically detect, enhance, catalog, and classify photographs and fingerprints from crime scenes with minimal human intervention, meeting or exceeding the 50% time-saving threshold for routine indexing and digitization workflows.
Task automatabilityclaude-sonnet-52/5Physical processing of film and prints (development, chain-of-custody handling, cataloging) is largely a manual/chemical or procedural workflow that current AI cannot execute end-to-end, though digital image organization could be assisted.:
Adoption barriersclaude-haiku-4-5-202510013/5Chain-of-custody requirements and potential evidentiary standards mean police departments often require human review and sign-off on automated processing, and there is organizational reluctance to fully displace officers, creating moderate adoption friction.
Adoption barriersclaude-sonnet-54/5Evidence handling requires strict chain-of-custody, certification, and legal admissibility standards, creating strong procedural and liability barriers to automation.
Cost vs. human wageclaude-haiku-4-5-202510014/5AI-driven scanning, enhancement, and classification infrastructure costs are substantially lower than the fully-loaded hourly rate of skilled police records officers when amortized across high-volume processing, approaching an order of magnitude savings at scale.
Cost vs. human wageclaude-sonnet-51/5AI cannot substitute for the physical chemical/digital processing equipment and chain-of-custody procedures, so no meaningful cost comparison favors AI.
Technical feasibility todayclaude-haiku-4-5-202510014/5Deployed products in forensic imaging (automated fingerprint identification systems, digital image enhancement software) and document processing work reliably in production, though some edge cases (damaged prints, unusual angles) still require human review, preventing a perfect 5.
Technical feasibility todayclaude-sonnet-51/5No deployed AI product performs physical film processing or forensic print handling in production; this remains a manual lab/technician task.

Photograph crime or accident scenes for evidence records.

43

CI 25–60 · exposure 45 · augmentation 63 · importance 4.6/5 · click for rater detail

Panel ratings per dimension for this task
DimensionModelRatingRationale
Sector adoption velocityclaude-haiku-4-5-202510012/5Law enforcement agencies operate with slow digital transformation compared to commercial sectors; bureaucratic procurement, liability caution, and union-related concerns limit rapid rollout. Pilot programs exist but production displacement remains very limited in most jurisdictions.
Sector adoption velocityclaude-sonnet-52/5Law enforcement is a traditionally slow-adopting sector for new technology due to budget constraints, procurement processes, and evidentiary standards, though drone use is slowly growing.
Augmentation potentialclaude-haiku-4-5-202510014/5AI imaging systems can dramatically augment officer productivity by automating 360-degree scene capture, generating structured evidence catalogs, and flagging anomalies for human review. This keeps the officer in the loop while freeing them from tedious manual photography and organization tasks.
Augmentation potentialclaude-sonnet-53/5AI can assist with scene reconstruction, image enhancement, and cataloging photos, helping officers organize and analyze evidence more efficiently even though the human still directs the capture process.
Task automatabilityclaude-haiku-4-5-202510014/5Current computer vision systems can autonomously capture, catalog, and archive photographs of static scenes with metadata (timestamps, GPS, evidence tagging) at high speed. A human officer could still oversee framing and evidence selection, but the capture, processing, and storage workflow is substantially automatable by deployed mobile/drone imaging agents with 60%+ time savings.
Task automatabilityclaude-sonnet-52/5While AI-enabled cameras and drones can capture images, deciding what constitutes probative evidence, ensuring chain-of-custody, and adapting to scene-specific conditions still requires trained human judgment on-site.'
Adoption barriersclaude-haiku-4-5-202510013/5Chain-of-custody and evidentiary standards require documented human oversight and sign-off; many jurisdictions mandate an officer review and certify scene documentation. Liability concerns around evidence admissibility and potential disputes over automated scene capture create adoption friction, though no hard legal prohibition exists yet.
Adoption barriersclaude-sonnet-54/5Evidence collection has strict chain-of-custody and authentication requirements, often requiring sworn officers to testify to how evidence was captured, creating strong legal/procedural barriers to full automation.
Cost vs. human wageclaude-haiku-4-5-202510014/5AI-driven imaging hardware (drones, automated multi-camera rigs, integration software) has declining per-incident costs compared to officer labor. Once systems are in place, marginal cost of capture and processing is low; human oversight costs remain but are substantially below full manual documentation.
Cost vs. human wageclaude-sonnet-52/5Specialized equipment (drones, LIDAR scanners) and required training add costs comparable to or exceeding officer time, especially given need for legal defensibility and equipment maintenance.
Technical feasibility todayclaude-haiku-4-5-202510013/5Autonomous evidence imaging systems exist in limited pilot deployments (drones, 360 cameras integrated with case management software), but widespread production adoption in law enforcement is still uncommon due to ecosystem fragmentation and department-specific workflows. Demonstrated capability is higher than research-stage, but reliable scale is not yet standard.
Technical feasibility todayclaude-sonnet-52/5Some departments use drones and 3D scanning tools for scene documentation, but these are narrow, specialized deployments rather than general AI systems reliably replacing officer photography.

Maintain records of evidence and write and review reports.

27

CI 20–34 · exposure 33 · augmentation 63 · importance 4.5/5 · click for rater detail

Panel ratings per dimension for this task
DimensionModelRatingRationale
Sector adoption velocityclaude-haiku-4-5-202510012/5Police departments adopt incremental digitization (case management, digital evidence logs) slowly due to budgetary constraints, legacy system inertia, and regulatory caution; AI-driven automation of records and reports remains rare in production.
Sector adoption velocityclaude-sonnet-52/5Law enforcement agencies are historically slow adopters of new digital systems due to budget constraints, legacy infrastructure, and cautious approaches to legally consequential documentation.
Augmentation potentialclaude-haiku-4-5-202510013/5AI can usefully assist by auto-populating evidence fields, suggesting report language, and flagging inconsistencies or missing data; this meaningfully speeds the officer's workflow while keeping them in control and accountable for accuracy.
Augmentation potentialclaude-sonnet-54/5AI can meaningfully assist by drafting report language, organizing evidence logs, and flagging inconsistencies, letting officers review and finalize rather than write from scratch.
Task automatabilityclaude-haiku-4-5-202510012/5AI can assist with evidence logging and routine report generation/review, but cannot reliably handle the full task end-to-end: judicial evidence chains require precise metadata, legal compliance, and human verification that AI cannot guarantee at 50% time-savings and equal quality.
Task automatabilityclaude-sonnet-53/5Report writing and evidence log maintenance involve structured, language-based work that AI drafting tools can significantly speed up, but chain-of-custody accuracy and legal admissibility require human verification, capping full end-to-end automation.
Adoption barriersclaude-haiku-4-5-202510015/5Police records systems are heavily regulated; evidence handling and chain-of-custody documentation have strict legal requirements, and any error carries liability. A sworn officer typically must sign off on evidence records and reports, creating a hard barrier to full automation.
Adoption barriersclaude-sonnet-54/5Evidence records are legally sensitive with strict chain-of-custody and admissibility requirements, often requiring sworn officer certification, creating strong regulatory and liability barriers to full automation.
Cost vs. human wageclaude-haiku-4-5-202510012/5AI systems for evidence management and report automation require substantial licensing, integration, and human oversight costs; they do not yet achieve cost parity with the loaded wage of a records officer, let alone savings.
Cost vs. human wageclaude-sonnet-52/5While AI drafting could cut time on report narratives, integration with secure chain-of-custody systems, audit requirements, and necessary human review keep overall costs closer to comparable rather than dramatically cheaper.
Technical feasibility todayclaude-haiku-4-5-202510012/5Products exist for evidence management (LIMS, case management systems) and automated report drafting, but real-world deployment shows material error rates in evidence classification, chain-of-custody accuracy, and legal sufficiency that require significant human oversight.
Technical feasibility todayclaude-sonnet-52/5AI transcription and report-drafting assistants exist in some police departments, but evidence records management systems are still largely manual-entry with human oversight, and no product reliably handles the full evidence-chain documentation task.

Create sketches and diagrams, by hand or computer software, to depict crime scenes.

25

CI 25–25 · exposure 25 · augmentation 50 · importance 3.6/5 · click for rater detail

Panel ratings per dimension for this task
DimensionModelRatingRationale
Sector adoption velocityclaude-haiku-4-5-202510012/5Law enforcement agencies adopt new technologies slowly, and crime scene documentation remains a process where human expertise and legal defensibility are paramount. Adoption of AI-generated crime scene sketches in production environments is minimal.
Sector adoption velocityclaude-sonnet-52/5Law enforcement agencies adopt digital tools slowly due to budget constraints, evidentiary standards, and conservative institutional culture, though 3D scanning tools are gradually gaining use.
Augmentation potentialclaude-haiku-4-5-202510013/5AI can meaningfully assist by auto-generating preliminary layouts, suggesting measurement scales, or providing template-based diagram structures that officers then refine and certify. This augmentation could accelerate the sketch creation process while maintaining human accountability.
Augmentation potentialclaude-sonnet-53/5Software tools (CAD, 3D laser scanning, photogrammetry) meaningfully speed up diagram creation and improve precision, while the officer still directs and validates the output.
Task automatabilityclaude-haiku-4-5-202510012/5AI can assist with diagram generation and layout, but crime scene sketches require careful spatial reasoning, accurate measurement integration, and contextual interpretation of evidence that varies significantly by scene. Current systems cannot reliably produce legally defensible, precise crime scene documentation without substantial human oversight.
Task automatabilityclaude-sonnet-52/5AI can assist with generating diagrams from measurements or photos, but accurately depicting a specific crime scene requires precise on-site data capture and forensic judgment that current tools cannot fully automate end-to-end.
Adoption barriersclaude-haiku-4-5-202510014/5Crime scene documentation must meet strict evidentiary standards, potentially requiring certification or legal sign-off by authorized personnel. Courts may require human-created or human-verified records, and accuracy errors carry high liability costs, creating substantial barriers to automation.
Adoption barriersclaude-sonnet-54/5Crime scene documentation must be legally defensible and chain-of-custody compliant, typically requiring a trained, often sworn officer to create and certify the record for court admissibility.
Cost vs. human wageclaude-haiku-4-5-202510012/5The combination of AI tools, required human verification, corrections, and liability oversight means total cost remains comparable to or potentially exceeds having trained personnel create sketches directly.
Cost vs. human wageclaude-sonnet-52/5Specialized forensic scanning/diagramming equipment and software have real costs comparable to or exceeding officer time, since human verification and on-site work are still required.
Technical feasibility todayclaude-haiku-4-5-202510012/5While image generation and diagramming tools exist, they lack the specialized capability to create forensically accurate crime scene sketches that meet law enforcement documentation standards. No mainstream deployed product reliably performs this task at the quality required for court presentation.
Technical feasibility todayclaude-sonnet-52/5CAD and forensic diagramming software (e.g., 3D scanning, laser measurement tools) exist and are used, but full AI-generated crime scene sketches from raw scene data are not yet a standard deployed product.

Coordinate or conduct instructional classes or in-services, such as citizen police academy classes and crime scene training for other officers.

13

CI 5–21 · exposure 13 · augmentation 50 · importance 3.4/5 · click for rater detail

Panel ratings per dimension for this task
DimensionModelRatingRationale
Sector adoption velocityclaude-haiku-4-5-202510012/5Law enforcement remains a traditional, human-centered sector with strict compliance and training protocols; while some agencies experiment with online modules, live instruction by qualified officers remains the standard practice.
Sector adoption velocityclaude-sonnet-51/5Law enforcement training programs are a low-digitization, tradition-bound sector with minimal AI adoption for instructional delivery.
Augmentation potentialclaude-haiku-4-5-202510013/5AI can assist officers in preparing instructional materials, generating scenario examples, or providing supplementary online content, but the core task of delivering and coordinating live training remains primarily human-driven.
Augmentation potentialclaude-sonnet-53/5AI can help generate training materials, quizzes, presentation slides, and even simulate scenarios, providing moderate assistance to the human instructor's preparation work.
Task automatabilityclaude-haiku-4-5-202510011/5This task requires real-time instruction delivery, managing classroom dynamics, responding to live questions, and adjusting teaching based on audience engagement—all of which demand human presence and judgment that current AI cannot replicate end-to-end in a classroom setting.
Task automatabilityclaude-sonnet-52/5Live instructional delivery, hands-on crime scene demonstrations, and coordination of citizen academy logistics require in-person presence, physical demonstration, and interactive facilitation that current AI cannot replace end-to-end.atability.pdf.title-attribute-remove notes: AI could help draft materials but not run the actual class.pptx.pptx.pptx.pptx.pptx.pptx.pptx.pptx.pptx.pptx.pptx.pptx.pptx.pptx.pptx.pptx.pptx.pptx.pptx.pptx.pptx.pptx.pptx.pptx.pptx.pptx.pptx.pptx.pptx.pptx.pptx.pptx.pptx.pptx.pptx.pptx.pptx.pptx.pptx.pptx.pptx.pptx.pptx.pptx.pptx.pptx.pptx.pptx.pptx.pptx.pptx.pptx.pptx.pptx.pptx.pptx.pptx.pptx.pptx.pptx.pptx.pptx.pptx.pptx.pptx.pptx.pptx.pptx.pptx.pptx.pptx.pptx.pptx.pptx.pptx.pptx.pptx.pptx.pptx.pptx.pptx.pptx.pptx.pptx.pptx.pptx.pptx.pptx.pptx.pptx.pptx.pptx.pptx.pptx.pptx.pptx.pptx.pptx.pptx.pptx.pptx.pptx.pptx.pptx.pptx.pptx.pptx
Adoption barriersclaude-haiku-4-5-202510014/5Law enforcement training, especially for crime scenes and police procedures, typically requires certification and accountability to training standards; instructors are often required to be credentialed officers, creating significant legal and regulatory barriers to full automation.
Adoption barriersclaude-sonnet-54/5Training officers and citizens in policing procedures typically requires sworn/certified personnel with institutional authority and credibility, creating strong organizational and credibility barriers to full automation.
Cost vs. human wageclaude-haiku-4-5-202510011/5The cost of an AI system capable of delivering credible law enforcement training, including necessary oversight and validation by subject-matter experts, would exceed the loaded wage of an experienced officer-instructor.
Cost vs. human wageclaude-sonnet-51/5Human instructors and coordinators are essential for scheduling, facilitation, and hands-on demonstration; AI cannot substitute at lower cost for the core delivery.
Technical feasibility todayclaude-haiku-4-5-202510012/5While AI can help generate training materials or draft lesson plans, no deployed product reliably conducts live instructional classes or manages in-service training as a primary instructor at scale in law enforcement contexts.
Technical feasibility todayclaude-sonnet-51/5No deployed product conducts or coordinates in-person police training or citizen academy classes; this remains a human-led organizational and pedagogical activity.

Package, store and retrieve evidence.

13

CI 0–25 · exposure 17 · augmentation 50 · importance 4.2/5 · click for rater detail

Panel ratings per dimension for this task
DimensionModelRatingRationale
Sector adoption velocityclaude-haiku-4-5-202510012/5Police departments are traditionally slow to adopt new technologies due to budget constraints, institutional conservatism, and tight regulation. While some large departments have implemented evidence management automation, overall sector adoption remains limited and fragmented.
Sector adoption velocityclaude-sonnet-51/5Law enforcement evidence handling is a physical, low-digitization process with minimal AI adoption for the physical handling components.
Augmentation potentialclaude-haiku-4-5-202510014/5AI-powered inventory systems, barcode/RFID tracking, and computer vision for evidence identification significantly assist officers in locating and organizing evidence, reducing manual search time and improving accuracy while the officer remains responsible for physical handling and legal compliance.
Augmentation potentialclaude-sonnet-52/5AI can assist with digital logging, barcoding systems, or database retrieval of records about evidence, but offers little help with the physical packaging and storage itself.
Task automatabilityclaude-haiku-4-5-202510012/5While AI can assist with cataloging and retrieval through computer vision and database systems, the task involves physical handling, chain-of-custody verification, and legal compliance that require human oversight. Automated storage/retrieval systems exist but do not meet the 50% time-saving bar for the full task including security, evidence integrity, and legal documentation.
Task automatabilityclaude-sonnet-51/5This is a physical chain-of-custody task involving handling, labeling, and securely storing physical evidence, which requires manual dexterity and physical presence that current AI cannot perform.'
Adoption barriersclaude-haiku-4-5-202510015/5Legal admissibility of evidence depends on strict chain-of-custody documentation and human accountability; evidence handling is often explicitly governed by state/local law and court rules requiring certified personnel. Liability for evidence loss or contamination creates hard barriers to full automation without human sign-off.
Adoption barriersclaude-sonnet-55/5Chain-of-custody and evidentiary integrity rules impose strict legal requirements that authorized, accountable personnel handle evidence, creating hard regulatory and liability barriers.
Cost vs. human wageclaude-haiku-4-5-202510012/5Initial capital investment in automated storage systems is high, and ongoing maintenance, monitoring, and human oversight remain necessary costs. The all-in cost is comparable to or slightly higher than human evidence technicians for many jurisdictions.
Cost vs. human wageclaude-sonnet-51/5There is no AI substitute performing this physical task, so cost comparison favors the human by default since no viable AI alternative exists.
Technical feasibility todayclaude-haiku-4-5-202510013/5Automated evidence storage and retrieval systems are deployed in some police departments, and barcode/RFID tracking is mature, but integration with chain-of-custody requirements and human verification remains necessary. Products exist but material gaps remain in autonomous end-to-end evidence handling without human involvement.
Technical feasibility todayclaude-sonnet-51/5No deployed AI product physically packages or retrieves physical evidence; this remains entirely a human/robotic manual task outside current AI product scope.

Look for trace evidence, such as fingerprints, hairs, fibers, or shoe impressions, using alternative light sources when necessary.

10

CI 0–20 · exposure 13 · augmentation 38 · importance 4.3/5 · click for rater detail

Panel ratings per dimension for this task
DimensionModelRatingRationale
Sector adoption velocityclaude-haiku-4-5-202510012/5Forensic agencies are conservative adopters due to legal constraints; while some labs use AI-assisted fingerprint matching and image analysis, widespread production deployment of AI in trace evidence collection remains limited and cautious.
Sector adoption velocityclaude-sonnet-51/5Law enforcement forensic fieldwork is a low-digitization, physically embedded sector with minimal AI deployment for on-scene evidence searching.
Augmentation potentialclaude-haiku-4-5-202510013/5AI can assist forensic officers by enhancing images, flagging potential matches, and automating database comparisons, moderately raising productivity in evidence review and pattern recognition phases of the work.
Augmentation potentialclaude-sonnet-52/5AI-enhanced imaging or pattern-matching tools can assist in analyzing collected evidence afterward, but offer little help during the physical search and detection phase itself.
Task automatabilityclaude-haiku-4-5-202510012/5Current AI systems can assist with image analysis of fingerprints and some pattern recognition on trace evidence photographs, but the task requires physical collection, proper handling of evidence, and judgment about what constitutes relevant trace material in context—capabilities that remain firmly human-dependent today.
Task automatabilityclaude-sonnet-51/5Physical evidence detection at crime scenes requires hands-on manipulation, mobility, and situational judgment in unpredictable environments that current AI cannot perform end-to-end.atable systems exist only for narrow image analysis, not the physical search process itself.
Adoption barriersclaude-haiku-4-5-202510015/5Chain of custody, admissibility of evidence in court, and legal requirements for forensic examination mandate human responsibility and sign-off; regulatory and liability frameworks make autonomous or unsupervised AI evidence collection legally impermissible.
Adoption barriersclaude-sonnet-55/5Chain-of-custody, evidentiary admissibility, and certification requirements mean only trained, often licensed personnel can legally collect and document forensic evidence for court use.
Cost vs. human wageclaude-haiku-4-5-202510012/5Forensic AI software is expensive, requires specialized hardware (alternative light sources, imaging equipment), trained operators, and significant infrastructure; the all-in cost per examination likely exceeds that of a trained forensic technician.
Cost vs. human wageclaude-sonnet-51/5There is no AI system substituting for the physical scene search, so cost comparison favors the human as the only viable option today.
Technical feasibility todayclaude-haiku-4-5-202510012/5While forensic image analysis tools exist and some labs use AI for fingerprint matching, no deployed system reliably performs end-to-end trace evidence detection and collection autonomously; systems require human operators to position, capture, and interpret findings.
Technical feasibility todayclaude-sonnet-51/5No deployed product autonomously searches physical scenes for trace evidence using alternative light sources; this remains a manual forensic task performed by trained officers.

Submit evidence to supervisors, crime labs, or court officials for legal proceedings.

4

CI 0–9 · exposure 8 · augmentation 38 · importance 4.4/5 · click for rater detail

Panel ratings per dimension for this task
DimensionModelRatingRationale
Sector adoption velocityclaude-haiku-4-5-202510011/5Law enforcement remains a low-digitization, highly regulated sector with strong legal and human-contact requirements. Adoption of AI for core evidence handling is minimal; most jurisdictions still require paper trails and human accountability.
Sector adoption velocityclaude-sonnet-51/5Law enforcement evidence handling is a highly regulated, low-digitization physical process with minimal AI adoption for this specific task.
Augmentation potentialclaude-haiku-4-5-202510012/5AI could assist by auto-populating forms, flagging missing documentation, or organizing evidence metadata, but the critical judgment call and legal authority still reside with the human officer, limiting transformative augmentation potential.
Augmentation potentialclaude-sonnet-53/5AI can assist with generating evidence logs, tracking chain-of-custody records, and drafting submission paperwork, improving efficiency around the core physical task.
Task automatabilityclaude-haiku-4-5-202510011/5Submitting evidence for legal proceedings requires human judgment about legal chain-of-custody, proper documentation, regulatory compliance, and court-specific requirements. AI cannot reliably determine what evidence goes where, to whom, and in what legal form without extensive human oversight that negates automation gains.
Task automatabilityclaude-sonnet-52/5The core action involves physical chain-of-custody handling, packaging, and legal submission of evidence, which requires physical presence and authorized signatures that AI cannot perform; only ancillary documentation/logging could be automated. "":""}
Adoption barriersclaude-haiku-4-5-202510015/5Legal and regulatory barriers are extremely high: chain-of-custody documentation, court rules of evidence, and liability for improper submission legally require human accountability and often explicit human sign-off by authorized personnel.
Adoption barriersclaude-sonnet-55/5Chain-of-custody and evidentiary admissibility rules require an authorized, accountable officer to physically handle and certify evidence transfer, making this a hard legal/procedural barrier.
Cost vs. human wageclaude-haiku-4-5-202510011/5The cost of AI infrastructure, integration with evidence management systems, legal review overhead, and mandatory human oversight would exceed the wage of a records officer performing the submission task directly.
Cost vs. human wageclaude-sonnet-51/5AI cannot substitute for the human physical and legal act of submitting evidence, so cost comparison favors the human performing the task entirely.
Technical feasibility todayclaude-haiku-4-5-202510011/5No deployed AI system performs end-to-end evidence submission for legal proceedings. This task involves legal accountability, regulatory compliance, and human sign-off that falls outside current AI product capabilities in production law enforcement systems.
Technical feasibility todayclaude-sonnet-51/5No deployed product handles physical evidence submission or legal chain-of-custody transfer to labs/courts; this remains a manual, procedurally regulated process.

Interview survivors, witnesses, suspects, and other law enforcement personnel.

3

CI 0–5 · exposure 5 · augmentation 50 · importance 3.8/5 · click for rater detail

Panel ratings per dimension for this task
DimensionModelRatingRationale
Sector adoption velocityclaude-haiku-4-5-202510011/5Law enforcement agencies have shown minimal adoption of AI for direct interviewing of witnesses or suspects; this task remains heavily human-dependent. No significant sector-wide trend toward AI-driven interviewing exists in police work.
Sector adoption velocityclaude-sonnet-51/5Law enforcement is a highly regulated, low-digitization sector for core investigative work, with AI adoption largely confined to back-office analytics rather than interviewing.
Augmentation potentialclaude-haiku-4-5-202510013/5AI can meaningfully assist officers by transcribing interviews, flagging inconsistencies for follow-up, and organizing notes, but the human officer must remain the primary interviewer. These tools improve workflow efficiency without displacing the human interviewer's core role.
Augmentation potentialclaude-sonnet-53/5AI can assist with transcription, translation, summarization, and flagging inconsistencies in statements, aiding officers without replacing the interview itself.
Task automatabilityclaude-haiku-4-5-202510011/5Interviewing survivors, witnesses, and suspects requires nuanced human judgment, empathy, trauma-informed responses, and real-time adaptation to emotional and safety contexts that current AI cannot reliably perform end-to-end. This task fundamentally depends on building rapport, reading non-verbal cues, and adjusting approach based on person-specific circumstances.
Task automatabilityclaude-sonnet-51/5Interviewing survivors, witnesses, and suspects requires real-time human judgment, rapport-building, adaptive questioning, and reading nonverbal cues that current AI cannot perform end-to-end in the field.
Adoption barriersclaude-haiku-4-5-202510015/5Interviewing suspects, witnesses, and victims involves legal requirements for human accountability, chain-of-custody integrity, Miranda rights administration, and liability for coerced or mishandled statements. Law enforcement and legal standards require a credentialed human officer to conduct and attest to interviews.
Adoption barriersclaude-sonnet-55/5Interviews of suspects and witnesses involve legal rights (Miranda warnings), evidentiary chain-of-custody, and officer authorization requirements that legally mandate a sworn human officer.
Cost vs. human wageclaude-haiku-4-5-202510011/5The human officer's loaded cost is relatively low for this critical law enforcement function, and AI systems capable of reliable independent interviewing do not exist at scale, making cost comparison moot. Even if such systems existed, liability and oversight costs would be substantial.
Cost vs. human wageclaude-sonnet-51/5There 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 todayclaude-haiku-4-5-202510012/5While AI can assist with transcription and note-taking during interviews, no deployed product reliably conducts interviews with vulnerable persons (survivors, suspects) independently. Research systems exist for limited interviewing contexts, but production deployment for law enforcement interviewing is not standard practice.
Technical feasibility todayclaude-sonnet-51/5No deployed product conducts law enforcement interviews autonomously; AI transcription and note-taking tools exist but the interview itself remains fully human-conducted.

Testify in court and present evidence.

0

CI 0–0 · exposure 0 · augmentation 25 · importance 4.4/5 · click for rater detail

Panel ratings per dimension for this task
DimensionModelRatingRationale
Sector adoption velocityclaude-haiku-4-5-202510011/5No meaningful AI adoption occurs for this task because it is legally and constitutionally barred; law enforcement and courts have no incentive or ability to replace human testimony with automation.
Sector adoption velocityclaude-sonnet-51/5Courts and legal proceedings are highly conservative, procedurally rigid, and have not adopted AI to replace witness testimony in any capacity.
Augmentation potentialclaude-haiku-4-5-202510012/5AI can modestly assist by helping officers organize evidence, draft summaries, or review case materials before testimony, but offers no augmentation during the actual courtroom presentation itself.
Augmentation potentialclaude-sonnet-52/5AI can help prepare testimony, organize evidence, or draft summaries beforehand, but offers minimal assistance during the actual act of testifying and presenting evidence in court.
Task automatabilityclaude-haiku-4-5-202510011/5Testifying in court and presenting evidence requires legal standing, cross-examination credibility, and real-time judgment in an adversarial setting—all fundamentally human roles. Current AI cannot appear as a witness or be held legally responsible for testimony.
Task automatabilityclaude-sonnet-51/5Testifying in court requires a real human witness with personal knowledge, credibility, and legal standing under oath; AI cannot perform this act at all, let alone with time savings.
Adoption barriersclaude-haiku-4-5-202510015/5Hard legal barriers: only a human with standing can testify under oath, and courts require the witness to be present and subject to cross-examination. Legislation and rules of evidence explicitly mandate human testimony.
Adoption barriersclaude-sonnet-55/5Testimony requires a sworn human witness subject to cross-examination, perjury laws, and rules of evidence; this is a hard legal barrier that categorically excludes AI substitution.
Cost vs. human wageclaude-haiku-4-5-202510011/5The task cannot be performed by AI at all, making cost comparison irrelevant; the only economic model is paying the required human officer or expert witness.
Cost vs. human wageclaude-sonnet-51/5There is no AI alternative to compare costs against since testimony must be given by a human witness under legal rules of evidence.
Technical feasibility todayclaude-haiku-4-5-202510011/5No deployed product can autonomously testify in court or present evidence as a witness. AI tools may assist in evidence organization or documentation, but the core task of courtroom testimony remains exclusively human.
Technical feasibility todayclaude-sonnet-51/5No deployed product performs courtroom testimony; this is fundamentally a human legal function with no AI substitute in production or research.

Dust selected areas of crime scene and lift latent fingerprints, adhering to proper preservation procedures.

0

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

Panel ratings per dimension for this task
DimensionModelRatingRationale
Sector adoption velocityclaude-haiku-4-5-202510011/5This is a core law enforcement function embedded in regulated criminal justice procedures. Adoption of AI in this context is minimal; the task remains performed by trained human specialists with no significant industry shift toward automation or AI agents.
Sector adoption velocityclaude-sonnet-51/5Physical forensic evidence collection in law enforcement is a low-digitization, hands-on field with essentially no AI/robotic adoption for this specific task.
Augmentation potentialclaude-haiku-4-5-202510012/5AI could potentially assist with record-keeping, image analysis of lifted prints, or guidance on best practices, but these are peripheral to the core physical task of dusting and lifting. The primary value remains with the skilled human technician.
Augmentation potentialclaude-sonnet-52/5AI can assist with related analysis (e.g., fingerprint matching software after collection) but offers minimal help with the physical dusting and lifting process itself.
Task automatabilityclaude-haiku-4-5-202510011/5Dusting and lifting latent fingerprints requires fine motor control, spatial judgment in crime scene conditions, and real-time decision-making about which surfaces to process. Current AI systems cannot physically manipulate crime scenes or make context-dependent preservation judgments that human experts perform.
Task automatabilityclaude-sonnet-51/5This is a manual, physical forensic task requiring hands-on dusting powder application and careful physical lifting of prints from surfaces at a physical crime scene; no AI system can perform the physical manipulation involved.
Adoption barriersclaude-haiku-4-5-202510015/5Fingerprint collection from crime scenes must be performed by trained and certified law enforcement personnel; legal and evidentiary standards mandate human collection and handling to ensure chain-of-custody and admissibility in court. Regulatory and liability frameworks require a human professional signature.
Adoption barriersclaude-sonnet-55/5Chain-of-custody, forensic evidentiary standards, and legal admissibility requirements mandate trained, certified personnel physically collect and preserve evidence, creating hard legal/procedural barriers.
Cost vs. human wageclaude-haiku-4-5-202510011/5The task requires specialized human training, proper equipment, and chain-of-custody protocols. No AI alternative exists that would reduce costs; if anything, hybrid human-AI approaches would add overhead rather than reduce it.
Cost vs. human wageclaude-sonnet-51/5There is no AI substitute performing this physical evidence-collection task, so cost comparison is moot—human labor is the only current option.
Technical feasibility todayclaude-haiku-4-5-202510011/5No deployed AI system performs the physical manipulation of dusting and lifting fingerprints from crime scenes. This is an inherently embodied task requiring specialized equipment and environmental adaptation that remains beyond current robotic or AI capabilities in practice.
Technical feasibility todayclaude-sonnet-51/5No deployed AI or robotic product performs physical fingerprint dusting and lifting at crime scenes; this remains a purely manual forensic technician task.

Analyze and process evidence at crime scenes, during autopsies, or in the laboratory, wearing protective equipment and using powders and chemicals.

0

CI 0–0 · exposure 0 · augmentation 25 · importance 4.2/5 · click for rater detail

Panel ratings per dimension for this task
DimensionModelRatingRationale
Sector adoption velocityclaude-haiku-4-5-202510011/5Law enforcement agencies have minimal ability to adopt AI for physical evidence processing due to regulatory, liability, and chain-of-custody requirements; automation in this domain remains negligible.
Sector adoption velocityclaude-sonnet-51/5Forensic/crime scene work is a highly physical, low-digitization niche within law enforcement with minimal AI/robotic deployment in evidence collection to date.
Augmentation potentialclaude-haiku-4-5-202510012/5While AI-powered imaging analysis and database matching can assist with evidence comparison and documentation, the core task of physically analyzing and processing evidence remains human-dependent, providing limited augmentation value for the hands-on work itself.
Augmentation potentialclaude-sonnet-52/5AI can assist with some lab analysis interpretation (e.g., pattern matching in databases) but offers little help with the core physical collection and chemical processing steps described.
Task automatabilityclaude-haiku-4-5-202510011/5This task requires hands-on physical manipulation of evidence, protective equipment, and chemical/powder application in uncontrolled crime scene environments. Current AI systems cannot physically interact with or process physical evidence, making end-to-end automation infeasible.
Task automatabilityclaude-sonnet-51/5This is a physical, hands-on task requiring manual manipulation of physical evidence, chemicals, and powders at a physical scene; current AI cannot perform physical evidence collection or processing.'
Adoption barriersclaude-haiku-4-5-202510015/5Forensic evidence handling and analysis are strictly regulated; chain of custody, expert certification, and legal admissibility require licensed human professionals to conduct and sign off on all evidence processing. Liability and legal requirements create hard barriers to automation.
Adoption barriersclaude-sonnet-55/5Evidence handling requires certified/licensed personnel, strict chain-of-custody and legal admissibility standards, and liability for contamination or evidentiary integrity, making this a hard-barrier task.
Cost vs. human wageclaude-haiku-4-5-202510011/5AI has no capability to perform this task, so cost comparison is moot; the human cost is the baseline and cannot be undercut by unavailable automation.
Cost vs. human wageclaude-sonnet-51/5There is no AI substitute performing this physical task, so cost comparison favors the human by default since no AI alternative exists to price.
Technical feasibility todayclaude-haiku-4-5-202510011/5No deployed AI system can independently wear protective equipment, handle evidence, apply powders and chemicals, or perform forensic analysis in the field or laboratory. This remains a human-performed task requiring physical presence and direct manipulation.
Technical feasibility todayclaude-sonnet-51/5No deployed AI product physically collects or chemically processes forensic evidence at scenes or in labs; this remains entirely human-performed work requiring physical dexterity and presence.

Perform emergency work during off-hours.

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-4-5-202510011/5Law enforcement is a laggard sector for AI adoption in core operational tasks; emergency response automation faces legal, union, and public resistance. Adoption of AI for emergency dispatch support exists, but not for autonomous off-hours emergency work.
Sector adoption velocityclaude-sonnet-51/5Law enforcement field operations, especially emergency response, are among the least digitized and slowest-adopting environments for AI automation.
Augmentation potentialclaude-haiku-4-5-202510012/5AI can assist with limited tasks like data lookup or record retrieval during emergency response, but the core emergency work—decision-making, intervention, legal authority—cannot be meaningfully augmented by AI while the human stays responsible.
Augmentation potentialclaude-sonnet-52/5AI can assist with dispatch coordination, documentation templates, or rapid database lookups once on scene, but offers little help during the core emergency response itself.
Task automatabilityclaude-haiku-4-5-202510011/5Emergency response during off-hours requires real-time human decision-making, situational awareness, and legal authority that cannot be meaningfully automated by current AI systems. No current technology can autonomously handle the unpredictable nature of police emergencies or assume legal responsibility.
Task automatabilityclaude-sonnet-51/5Emergency off-hours work requires physical presence, scene assessment, chain-of-custody handling, and real-time judgment under unpredictable conditions that AI cannot perform end-to-end today.
Adoption barriersclaude-haiku-4-5-202510015/5Police authority and emergency response are strictly governed by law; only licensed, certified police officers can legally perform emergency work. Liability, legal accountability, and constitutional requirements create near-insurmountable barriers to automation.
Adoption barriersclaude-sonnet-55/5Sworn/certified personnel with legal authority, chain-of-custody responsibility, and liability for evidence handling are required, creating hard legal and organizational barriers to automation.
Cost vs. human wageclaude-haiku-4-5-202510011/5Emergency police work requires paid on-call officers or overtime wages, which remain cheaper than developing and maintaining AI systems capable of autonomous emergency response with liability coverage and 24/7 availability.
Cost vs. human wageclaude-sonnet-51/5There is no viable AI substitute performing this physical, on-call emergency task, so no meaningful cost comparison favors AI.
Technical feasibility todayclaude-haiku-4-5-202510011/5No deployed AI product reliably performs emergency police work autonomously. This task inherently requires human officers with legal authority, training, and accountability—AI systems cannot substitute for on-call police personnel.
Technical feasibility todayclaude-sonnet-51/5No deployed product performs emergency forensic/records response duties in the field; this remains squarely a human first-responder function.

Serve as technical advisor and coordinate with other law enforcement workers or legal personnel to exchange information on crime scene collection activities.

0

CI 0–0 · exposure 0 · augmentation 25 · importance 3.7/5 · click for rater detail

Panel ratings per dimension for this task
DimensionModelRatingRationale
Sector adoption velocityclaude-haiku-4-5-202510011/5Law enforcement remains a slow-adopting sector for automation, particularly in evidence handling and legal coordination. Institutional conservatism, regulatory compliance, and the human-critical nature of evidence integrity prevent rapid AI adoption here.
Sector adoption velocityclaude-sonnet-51/5Law enforcement crime scene operations are a low-digitization, physically grounded sector with minimal AI agent deployment for advisory coordination tasks.
Augmentation potentialclaude-haiku-4-5-202510012/5AI could provide minor assistance (e.g., summarizing crime scene protocols or flagging procedural checklists), but the core task of technical advising and inter-agency coordination relies heavily on human judgment, trust, and decision-making authority.
Augmentation potentialclaude-sonnet-52/5AI can help organize case notes, records, or communications logs, but offers limited assistance to the core interpersonal advisory and coordination function.
Task automatabilityclaude-haiku-4-5-202510011/5This task fundamentally requires human judgment, institutional knowledge, and direct coordination with law enforcement and legal personnel. Current AI cannot conduct real-time inter-agency communication, understand nuanced legal requirements, or serve as a trusted technical advisor in a chain-of-custody context.
Task automatabilityclaude-sonnet-51/5This requires live judgment-based interagency coordination, verbal advisory expertise, and real-time crime scene decision-making that AI cannot substitute for end-to-end.
Adoption barriersclaude-haiku-4-5-202510015/5Strong legal and regulatory barriers: officers must be licensed/certified, maintain chain of custody, and coordinate with legal personnel on matters requiring professional judgment and accountability. Liability for errors in crime scene evidence coordination creates hard substitution barriers.
Adoption barriersclaude-sonnet-55/5Chain-of-custody, legal admissibility, and law enforcement authorization requirements mean only certified personnel can serve in this advisory/coordination capacity.
Cost vs. human wageclaude-haiku-4-5-202510011/5A trained police identification officer with institutional credentials and legal liability is essential; any AI system would require constant human oversight and verification, making it more costly than direct human performance.
Cost vs. human wageclaude-sonnet-51/5There is no viable AI substitute performing this coordination role, so cost comparison favors the human specialist entirely.
Technical feasibility todayclaude-haiku-4-5-202510011/5No deployed AI system can reliably serve as a technical advisor or coordinate across law enforcement agencies and legal teams in production. This requires organizational authority, accountability, and domain expertise that current tools do not possess.
Technical feasibility todayclaude-sonnet-51/5No deployed AI product acts as a technical advisor coordinating live crime scene work between law enforcement and legal personnel; this remains a human-expert role.

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.