Forensic Science Technicians
19-4092.00Median wage $72,060/yr19,120 employed (US)Rank #827 of 923 scored · top 90% by substitution
Collect, identify, classify, and analyze physical evidence related to criminal investigations. Perform tests on weapons or substances, such as fiber, hair, and tissue to determine significance to investigation. May testify as expert witnesses on evidence or crime laboratory techniques. May serve as specialists in area of expertise, such as ballistics, fingerprinting, handwriting, or biochemistry.
Sub-scores
0–100 · band = confidence interval from rater disagreement
Substitution14
Exposure15
Augmentation46
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
24 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%16
panel mean rating 1.6/5 → substitution pressure 16/100
Technical feasibility todayw 20%13
panel mean rating 1.5/5 → substitution pressure 13/100
Cost vs. human wagew 15%14
panel mean rating 1.6/5 → substitution pressure 14/100
Adoption barriersw 20%inverted — strong barriers lower the score11
panel mean rating 4.6/5 (barrier strength) → substitution pressure 11/100
Sector adoption velocityw 10%12
panel mean rating 1.5/5 → substitution pressure 12/100
Task breakdown (24 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.
Keep records and prepare reports detailing findings, investigative methods, and laboratory techniques.
36CI 34–37 · exposure 41 · augmentation 63 · importance 4.8/5 · click for rater detail
Keep records and prepare reports detailing findings, investigative methods, and laboratory techniques.
36| Dimension | Model | Rating | Rationale |
|---|---|---|---|
| Sector adoption velocity | claude-haiku-4-5-20251001 | 2/5 | Forensic labs are moderately digitized but risk-averse; adoption of AI for official reports is slow due to legal and liability concerns. Most labs use templates and documentation tools, not generative AI for final reports. |
| Sector adoption velocity | claude-sonnet-5 | 2/5 | Forensic science labs are typically government-run, resource-constrained, and cautious about adopting AI tools for legally consequential documentation, resulting in slow uptake compared to information-sector norms. |
| Augmentation potential | claude-haiku-4-5-20251001 | 3/5 | AI can assist by auto-populating standard sections, organizing raw data into report structure, and drafting preliminary text, which would meaningfully speed technicians' work. The human remains responsible for accuracy and final sign-off, making augmentation plausible and valuable. |
| Augmentation potential | claude-sonnet-5 | 4/5 | AI can meaningfully assist by organizing data, generating report templates, and summarizing findings, boosting technician efficiency while the technician retains responsibility for accuracy and legal validity. |
| Task automatability | claude-haiku-4-5-20251001 | 3/5 | AI can automate parts of report generation—structuring findings, formatting data, and drafting boilerplate sections—but requires significant human review, verification of accuracy, and professional judgment about what findings are legally and scientifically material. This achieves partial time savings but not the ≥50% bar end-to-end at equal quality. |
| Task automatability | claude-sonnet-5 | 3/5 | AI can draft structured reports from lab data and notes, but forensic reports require precise, verifiable methodology descriptions and chain-of-custody accuracy that still need substantial human verification, limiting full end-to-end automation. |
| Adoption barriers | claude-haiku-4-5-20251001 | 4/5 | Forensic reports must often meet court admissibility standards, chain-of-custody requirements, and expert testimony standards under Daubert/FRE 702. A licensed technician's signature and professional accountability are typically legally required, creating hard barriers to full automation. |
| Adoption barriers | claude-sonnet-5 | 4/5 | Forensic reports often serve as legal evidence requiring certified technician sign-off, chain-of-custody integrity, and accountability under legal/regulatory scrutiny, creating strong barriers to full automation. |
| Cost vs. human wage | claude-haiku-4-5-20251001 | 2/5 | Integration, verification, legal review, and the cost of errors in forensic documentation are substantial. The loaded cost of a forensic technician is moderate, and AI + oversight overhead is not yet significantly cheaper than careful human authoring. |
| Cost vs. human wage | claude-sonnet-5 | 3/5 | AI drafting could cut some documentation time, but the need for expert review, accuracy verification, and legal defensibility keeps oversight costs meaningful, making savings moderate rather than order-of-magnitude. |
| Technical feasibility today | claude-haiku-4-5-20251001 | 2/5 | While tools exist for document automation and data structuring, no mature production system reliably generates forensic reports meeting legal and evidentiary standards without substantial human oversight. Error rates in automated legal/technical documentation remain high enough that deployment is limited to narrow, pre-vetted contexts. |
| Technical feasibility today | claude-sonnet-5 | 2/5 | General-purpose LLMs can generate report drafts, but no widely deployed forensic-specific product reliably produces court-admissible technical reports without extensive human review and validation. |
Use chemicals or other substances to examine latent fingerprint evidence and compare developed prints to those of known persons in databases.
28CI 25–30 · exposure 30 · augmentation 75 · importance 4.5/5 · click for rater detail
Use chemicals or other substances to examine latent fingerprint evidence and compare developed prints to those of known persons in databases.
28| Dimension | Model | Rating | Rationale |
|---|---|---|---|
| Sector adoption velocity | claude-haiku-4-5-20251001 | 2/5 | Forensic labs have adopted AFIS for database searching, but this is decades-old technology; labs remain heavily dependent on skilled human technicians for latent development and expert comparison, with little evidence of rapid displacement or end-to-end automation in practice. |
| Sector adoption velocity | claude-sonnet-5 | 2/5 | Forensic labs are typically public-sector, resource-constrained, and cautious about adopting new technology due to legal admissibility concerns, resulting in slow uptake of new AI tools beyond established AFIS systems. |
| Augmentation potential | claude-haiku-4-5-20251001 | 4/5 | AFIS candidates and automated comparison tools substantially assist technicians by narrowing search space and flagging potential matches, allowing experts to focus human judgment on verification and exclusion; this is a mature augmentation pattern in forensic practice. |
| Augmentation potential | claude-sonnet-5 | 4/5 | AFIS databases and image-enhancement software substantially speed up the comparison and candidate-matching portion of the task, meaningfully augmenting technician productivity even though final identification and lab work remain human-led. |
| Task automatability | claude-haiku-4-5-20251001 | 2/5 | Latent fingerprint development (chemical/substance application) involves physical manipulation that requires precision and environmental control; current robotics cannot reliably handle this. Print comparison has some automation potential through image-matching algorithms, but the nuanced judgment of partial or degraded prints, feature weighting, and exclusionary conclusions still requires expert human oversight. |
| Task automatability | claude-sonnet-5 | 2/5 | The chemical development of latent prints is a hands-on physical lab process requiring manipulation of physical evidence, which AI cannot perform; only the digital comparison/matching sub-step is automatable today.ed |
| Adoption barriers | claude-haiku-4-5-20251001 | 4/5 | Fingerprint evidence in criminal cases has strict chain-of-custody, admissibility, and expert-qualification requirements; courts typically require certified human examiners to testify, and liability for misidentification is high, creating legal and regulatory barriers to full automation. |
| Adoption barriers | claude-sonnet-5 | 4/5 | Forensic evidence handling requires certified technicians, strict chain-of-custody, courtroom admissibility standards, and legal accountability for testimony, creating strong regulatory and liability barriers to full automation. |
| Cost vs. human wage | claude-haiku-4-5-20251001 | 2/5 | AFIS systems and image-analysis software exist but require significant capital investment, maintenance, and ongoing expert technician labor for interpretation and quality assurance; the all-in cost is not yet substantially lower than skilled human technicians performing the work. |
| Cost vs. human wage | claude-sonnet-5 | 2/5 | Matching software is cheap per query, but the overall task including physical evidence handling, chain-of-custody, and chemical processing still requires paid technician labor, keeping blended costs comparable to human-only workflows. |
| Technical feasibility today | claude-haiku-4-5-20251001 | 2/5 | AFIS (Automated Fingerprint Identification System) performs database matching at scale, but human examiners remain essential for validation, exclusion determinations, and testimony; no deployed system reliably completes the full task (development + comparison + expert judgment) without expert human review. |
| Technical feasibility today | claude-sonnet-5 | 3/5 | AFIS and other automated fingerprint matching systems are mature and widely deployed for comparison, but the physical chemical development step has no deployed automation and still requires human forensic technicians. |
Determine types of bullets and specific weapons used in shootings.
23CI 20–25 · exposure 25 · augmentation 50 · importance 4.0/5 · click for rater detail
Determine types of bullets and specific weapons used in shootings.
23| Dimension | Model | Rating | Rationale |
|---|---|---|---|
| Sector adoption velocity | claude-haiku-4-5-20251001 | 2/5 | Adoption of ballistic AI is slow; most law enforcement still relies on manual expert review and comparison under microscopes. While some jurisdictions have invested in ballistic databases (e.g., NIBIN), AI-assisted analysis remains a pilot or nascent tool, not yet widespread in production. |
| Sector adoption velocity | claude-sonnet-5 | 2/5 | Forensic labs are historically slow to adopt new technology due to legal validation requirements, budget constraints, and accreditation processes, resulting in narrow, incremental tool adoption rather than broad AI deployment. |
| Augmentation potential | claude-haiku-4-5-20251001 | 3/5 | AI-assisted ballistic image analysis can help forensic technicians by pre-screening candidate matches, automating striations extraction, and ranking similar samples, reducing manual microscopy time. However, the final determination still requires expert human judgment, making this moderately augmentative rather than transformative. |
| Augmentation potential | claude-sonnet-5 | 3/5 | AI-assisted image matching and pattern recognition tools can help examiners narrow down candidate matches faster, improving throughput while the technician retains final judgment and testimony responsibility. |
| Task automatability | claude-haiku-4-5-20251001 | 2/5 | Ballistic comparison requires matching images of fired bullets and casings to known samples, which involves image analysis AI can partially automate (e.g., extracting microscopy images and comparing striations). However, the final determination of weapon association requires integration with case law, courtroom standards, and human expert judgment in ways that current AI cannot reliably replicate end-to-end with 50% time savings at equal quality. |
| Task automatability | claude-sonnet-5 | 2/5 | This requires physical examination of ballistic evidence, microscopy, and comparison against physical databases (e.g., NIBIN), which current general AI cannot perform end-to-end; only narrow sub-steps like image comparison assistance could be partially automated. |
| Adoption barriers | claude-haiku-4-5-20251001 | 4/5 | Forensic testimony on ballistic evidence is subject to Daubert standards and must meet rigorous courtroom evidentiary thresholds. Results must be reviewed and testified to by a licensed forensic expert, and errors carry high liability cost (wrongful conviction). These legal and professional gatekeeping requirements are hard barriers. |
| Adoption barriers | claude-sonnet-5 | 5/5 | Forensic conclusions used in criminal proceedings require certified, legally accountable examiners whose testimony can be challenged in court; liability, chain-of-custody, and admissibility standards make full automation legally untenable. |
| Cost vs. human wage | claude-haiku-4-5-20251001 | 2/5 | Ballistic analysis systems require specialized hardware (comparison microscopes, 3D imaging), expert training for interpretation, and substantial integration overhead. These costs remain substantial relative to paying a trained forensic technician, especially when accounting for validation and legal defensibility. |
| Cost vs. human wage | claude-sonnet-5 | 2/5 | Ballistic imaging systems and comparison microscopes are costly to acquire and maintain, and human examiner oversight is still required, so total cost is not dramatically lower than a trained technician's labor for this task. |
| Technical feasibility today | claude-haiku-4-5-20251001 | 2/5 | While ballistic imaging analysis tools exist in research and some law enforcement settings, they are narrow in scope and have not achieved production reliability for autonomous weapon identification. Current systems assist but do not replace the forensic expert; error rates and false positives remain too high for unassisted deployment. |
| Technical feasibility today | claude-sonnet-5 | 2/5 | Specialized ballistic imaging systems (e.g., IBIS/NIBIN) exist and are deployed, but they assist trained examiners rather than autonomously determining bullet/weapon type, and courtroom-grade conclusions still require human forensic expertise. |
Analyze data from computers or other digital media sources for evidence related to criminal activity.
23CI 20–25 · exposure 25 · augmentation 63 · importance 3.5/5 · click for rater detail
Analyze data from computers or other digital media sources for evidence related to criminal activity.
23| Dimension | Model | Rating | Rationale |
|---|---|---|---|
| Sector adoption velocity | claude-haiku-4-5-20251001 | 2/5 | Law enforcement and forensic labs are traditionally conservative, risk-averse sectors with strong procedural and legal constraints. While some digital tools are in use, meaningful AI-driven automation adoption remains minimal; most agencies still rely on trained human examiners. |
| Sector adoption velocity | claude-sonnet-5 | 2/5 | Law enforcement and forensic labs are historically slow technology adopters due to budget constraints, procurement processes, and evidentiary standards, resulting in limited production AI deployment. |
| Augmentation potential | claude-haiku-4-5-20251001 | 3/5 | AI can assist forensic examiners by accelerating data indexing, flagging suspicious patterns, and reducing time spent on routine file classification. However, the human examiner must interpret findings and make legally defensible conclusions, limiting the transformative potential of current tools. |
| Augmentation potential | claude-sonnet-5 | 4/5 | AI tools significantly speed up data triage, indexing, pattern detection, and anomaly flagging within large digital datasets, meaningfully boosting technician throughput while they retain final analytical control. |
| Task automatability | claude-haiku-4-5-20251001 | 2/5 | While AI can assist with pattern detection and keyword searching in large datasets, forensic analysis requires chain-of-custody integrity, legal interpretation of evidence, and judgment about relevance that humans must oversee. Current AI cannot reliably perform the full investigative reasoning without significant human validation, falling well short of 50% time savings at equal quality. |
| Task automatability | claude-sonnet-5 | 2/5 | Digital forensics analysis requires interpreting complex, case-specific evidence chains, legal defensibility, and judgment about intent/context that current AI cannot reliably replicate end-to-end.'}, |
| Adoption barriers | claude-haiku-4-5-20251001 | 5/5 | Legal and regulatory frameworks (evidence admissibility rules, chain of custody, expert testimony requirements) mandate human forensic examiners and typically require certification/licensing. Courts require human experts to authenticate and defend findings, creating hard barriers to full automation. |
| Adoption barriers | claude-sonnet-5 | 4/5 | Chain-of-custody rules, admissibility standards, and certification requirements mean a qualified human technician must typically perform and attest to the analysis for legal proceedings. |
| Cost vs. human wage | claude-haiku-4-5-20251001 | 2/5 | Forensic software licenses and integration are expensive, and oversight by certified forensic examiners is mandatory. The per-analysis cost remains comparable to or exceeds the human examiner wage, especially accounting for liability and quality assurance requirements. |
| Cost vs. human wage | claude-sonnet-5 | 2/5 | Specialized forensic AI tools require significant licensing, integration, and expert oversight, so total cost is not dramatically lower than trained technician labor for this task. |
| Technical feasibility today | claude-haiku-4-5-20251001 | 2/5 | Tools exist for digital evidence extraction and basic analysis (e.g., EnCase, Forensic Toolkit), but these are narrow in scope and still require forensic examiner expertise. No deployed AI system reliably performs end-to-end forensic analysis independently; deployed products handle specific subtasks under human control. |
| Technical feasibility today | claude-sonnet-5 | 2/5 | AI-assisted tools exist for keyword search, file carving, and pattern flagging, but comprehensive forensic analysis for court-admissible evidence remains human-led in production settings. |
Measure and sketch crime scenes to document evidence.
21CI 18–25 · exposure 25 · augmentation 50 · importance 4.4/5 · click for rater detail
Measure and sketch crime scenes to document evidence.
21| Dimension | Model | Rating | Rationale |
|---|---|---|---|
| Sector adoption velocity | claude-haiku-4-5-20251001 | 1/5 | Forensic labs are highly regulated, conservative, and typically public-sector agencies with slow technology adoption. Liability and evidentiary standards create structural resistance to delegating scene documentation to automated systems. |
| Sector adoption velocity | claude-sonnet-5 | 2/5 | Law enforcement and forensic science are traditionally slow-adopting sectors with budget constraints, though some larger agencies have adopted 3D scanning tools over the past decade. |
| Augmentation potential | claude-haiku-4-5-20251001 | 3/5 | AI-assisted measurement alignment, automated diagram templates, and evidence-highlighting overlays could modestly speed up technician workflow, but the human must remain in control to ensure legal admissibility and accuracy. |
| Augmentation potential | claude-sonnet-5 | 3/5 | Photogrammetry and 3D scanning tools meaningfully speed up documentation and improve accuracy, but a trained technician must still operate equipment, verify results, and exercise judgment on scene. |
| Task automatability | claude-haiku-4-5-20251001 | 2/5 | While AI can assist with image analysis and diagram generation, the task requires spatial judgment, selective documentation of legally relevant evidence, and chain-of-custody accountability. Current systems cannot reliably measure, decide what matters, and sketch crime scenes end-to-end without human oversight. |
| Task automatability | claude-sonnet-5 | 2/5 | AI/robotic tools can assist with 3D scanning and photogrammetry-based sketching, but the physical measurement, evidence identification, and on-scene judgment required cannot be fully automated end-to-end today. |
| Adoption barriers | claude-haiku-4-5-20251001 | 5/5 | Forensic documentation is legally mandated and forms evidence for criminal proceedings; a licensed technician must personally perform and attest to measurements and sketches. Admissibility and chain-of-custody rules require human accountability that AI cannot satisfy. |
| Adoption barriers | claude-sonnet-5 | 4/5 | Crime scene documentation is subject to strict chain-of-custody, evidentiary admissibility standards, and legal requirements that typically mandate certified technicians perform or validate this work. |
| Cost vs. human wage | claude-haiku-4-5-20251001 | 2/5 | Specialized measurement hardware (LiDAR, drones) and human technician oversight still dominate costs. AI tooling for evidence identification and sketching would add cost and complexity without replacing the technician, making it costlier than current practice. |
| Cost vs. human wage | claude-sonnet-5 | 2/5 | 3D scanning equipment and software represent significant capital investment and training costs that often exceed the marginal cost of a technician manually sketching, especially for smaller departments. |
| Technical feasibility today | claude-haiku-4-5-20251001 | 2/5 | No deployed forensic product performs this task autonomously. LiDAR and photogrammetry tools exist for measurement but require human-directed setup; AI sketch generation is nascent and not used in forensic practice at scale. |
| Technical feasibility today | claude-sonnet-5 | 2/5 | Products like laser scanners and photogrammetry software (e.g., FARO, Leica) exist and are used by some agencies, but they require trained operators and are not universally deployed or fully autonomous. |
Interpret laboratory findings or test results to identify and classify substances, materials, or other evidence collected at crime scenes.
20CI 20–20 · exposure 25 · augmentation 50 · importance 4.0/5 · click for rater detail
Interpret laboratory findings or test results to identify and classify substances, materials, or other evidence collected at crime scenes.
20| Dimension | Model | Rating | Rationale |
|---|---|---|---|
| Sector adoption velocity | claude-haiku-4-5-20251001 | 2/5 | Forensic labs are conservative institutions bound by accreditation standards (ASCLD-LAB, ISO 17025) and legal requirements; adoption of AI-driven interpretation is slow and limited to narrow, well-validated use cases rather than broad deployment. |
| Sector adoption velocity | claude-sonnet-5 | 2/5 | Forensic science is a specialized, tightly regulated public-sector field with slow technology adoption cycles, heavy reliance on validated legacy methods, and cautious integration of new tools due to legal scrutiny. |
| Augmentation potential | claude-haiku-4-5-20251001 | 3/5 | AI can meaningfully assist forensic technicians by automating initial screening, pattern detection in large datasets, and hypothesis generation, but the technician must retain final interpretive authority and judgment for legal and evidentiary integrity. |
| Augmentation potential | claude-sonnet-5 | 3/5 | AI-assisted pattern recognition, database searching, and data visualization can meaningfully speed up preliminary analysis and hypothesis generation, though the technician still performs and certifies final interpretation. |
| Task automatability | claude-haiku-4-5-20251001 | 2/5 | While AI can assist in pattern matching and spectral analysis, forensic interpretation requires expert judgment on complex evidentiary chains, legal admissibility, and situational context that goes beyond current automated systems. Significant human oversight and decision-making remain necessary for reliable results. |
| Task automatability | claude-sonnet-5 | 2/5 | AI can assist with pattern matching and data analysis (e.g., spectral library matching) but final interpretation of forensic evidence requires expert judgment, chain-of-custody accountability, and courtroom-defensible reasoning that current systems cannot fully replicate end-to-end. |
| Adoption barriers | claude-haiku-4-5-20251001 | 5/5 | Forensic evidence interpretation is heavily regulated and must meet strict chain-of-custody, admissibility, and expert testimony standards in legal proceedings; human forensic technicians and expert witnesses are often legally required to authenticate and defend findings in court. |
| Adoption barriers | claude-sonnet-5 | 5/5 | Forensic interpretation is subject to strict legal and evidentiary standards, requiring certified/licensed forensic scientists to testify and certify results, creating hard regulatory and liability barriers to automation. |
| Cost vs. human wage | claude-haiku-4-5-20251001 | 2/5 | While analytical instruments provide efficiency gains, the specialized training, licensing, and liability costs for forensic technicians remain substantial; AI tools still require significant human oversight and domain expertise, keeping total cost comparable to or higher than human-only workflows. |
| Cost vs. human wage | claude-sonnet-5 | 2/5 | Specialized forensic analysis software and instrumentation carry high costs, and the need for expert verification to ensure legal admissibility means AI does not yet provide a clear cost advantage over trained technicians. |
| Technical feasibility today | claude-haiku-4-5-20251001 | 2/5 | Some narrow AI tools exist for specific analytical tasks (e.g., fingerprint or DNA sequence matching), but no deployed system reliably interprets the full range of forensic evidence types or produces defensible courtroom-ready conclusions without expert human validation. |
| Technical feasibility today | claude-sonnet-5 | 2/5 | Some lab software uses automated matching for substances (drug identification via spectroscopy databases) but comprehensive interpretation combining multiple evidence types into classifications is not yet handled reliably by deployed products without expert review. |
Identify and quantify drugs or poisons found in biological fluids or tissues, in foods, or at crime scenes.
20CI 20–20 · exposure 25 · augmentation 50 · importance 3.7/5 · click for rater detail
Identify and quantify drugs or poisons found in biological fluids or tissues, in foods, or at crime scenes.
20| Dimension | Model | Rating | Rationale |
|---|---|---|---|
| Sector adoption velocity | claude-haiku-4-5-20251001 | 2/5 | Forensic laboratories are government or institutional entities with slow modernization cycles and strong regulatory constraints. While some labs adopt AI-assisted analysis tools, deep automation remains minimal due to legal requirements and the specialized, regulated nature of forensic work. |
| Sector adoption velocity | claude-sonnet-5 | 2/5 | Forensic science is a specialized, heavily regulated public-sector field with slow technology adoption cycles and limited AI integration into accredited lab workflows. |
| Augmentation potential | claude-haiku-4-5-20251001 | 3/5 | AI tools can meaningfully assist forensic technicians by automating spectral peak identification, database matching, and statistical analysis of toxicology data, reducing manual review time. However, the human expert must retain control over interpretation and validation for legal and safety reasons. |
| Augmentation potential | claude-sonnet-5 | 3/5 | AI-assisted spectral libraries, pattern recognition, and data analysis software can help technicians identify compounds faster, though the technician must verify and interpret results manually. |
| Task automatability | claude-haiku-4-5-20251001 | 2/5 | Current AI can assist in data analysis and pattern matching of toxicology results, but cannot independently perform the complex wet-chemistry analysis, sample preparation, instrumentation operation, and chain-of-custody documentation required. The task requires hands-on laboratory work and instrumental expertise that remains firmly in human domain. |
| Task automatability | claude-sonnet-5 | 2/5 | The physical analytical work (sample prep, chromatography, mass spec, chain-of-custody) requires laboratory instrumentation and hands-on technique that AI cannot perform end-to-end; AI can assist in interpreting spectral data but not replace the full workflow.' |
| Adoption barriers | claude-haiku-4-5-20251001 | 5/5 | Forensic toxicology findings must be generated by certified personnel and are subject to strict chain-of-custody, legal admissibility standards (Daubert), and regulatory oversight by agencies like ASCLD and SWGDRUG. Courts require human expert testimony and sign-off, creating hard legal barriers to full automation. |
| Adoption barriers | claude-sonnet-5 | 5/5 | Forensic toxicology results must be produced and certified by qualified, often licensed technicianss/analysts under strict chain-of-custody and evidentiary admissibility rules, with legal liability for errors in criminal proceedings. |
| Cost vs. human wage | claude-haiku-4-5-20251001 | 2/5 | Forensic toxicology demands high-precision laboratory equipment, certified technicians, and rigorous quality controls that remain expensive. AI data-analysis tools reduce some overhead but cannot eliminate the substantial labor and capital costs of forensic chemistry, keeping total cost comparable to or exceeding human-only workflows. |
| Cost vs. human wage | claude-sonnet-5 | 2/5 | Instrumentation, reagents, and technician oversight remain necessary regardless of AI use, so cost savings are limited to data analysis portions, not the full task cost structure. |
| Technical feasibility today | claude-haiku-4-5-20251001 | 2/5 | While AI-powered chromatography software and spectral analysis tools exist to support toxicology workflows, no deployed product autonomously identifies and quantifies unknown substances end-to-end. Current systems require expert human chemists to design protocols, interpret results, and validate findings for legal admissibility. |
| Technical feasibility today | claude-sonnet-5 | 2/5 | Deployed AI tools exist for spectral pattern matching and data interpretation in toxicology labs, but no product autonomously performs full sample collection, testing, and legally defensible identification/quantification today. |
Use photographic or video equipment to document evidence or crime scenes.
19CI 14–25 · exposure 20 · augmentation 50 · importance 4.7/5 · click for rater detail
Use photographic or video equipment to document evidence or crime scenes.
19| Dimension | Model | Rating | Rationale |
|---|---|---|---|
| Sector adoption velocity | claude-haiku-4-5-20251001 | 2/5 | Forensic labs and law enforcement are risk-averse and heavily regulated; adoption of autonomous or minimally supervised documentation remains slow. Most agencies still rely on trained technicians for on-scene capture despite advances in imaging technology. |
| Sector adoption velocity | claude-sonnet-5 | 1/5 | Forensic science and law enforcement are slow-adopting, highly procedural, physically-grounded sectors with minimal AI-driven automation of on-scene physical documentation tasks. |
| Augmentation potential | claude-haiku-4-5-20251001 | 3/5 | AI can assist with post-capture tasks (image enhancement, cataloging, angle suggestions) and real-time feedback on lighting or focus, but the core decision-making about what to document remains primarily human-driven in current practice. |
| Augmentation potential | claude-sonnet-5 | 3/5 | AI can assist with post-capture tasks like image enhancement, cataloging, tagging, or generating scene diagrams from photos, improving efficiency even though the core capture remains manual. |
| Task automatability | claude-haiku-4-5-20251001 | 2/5 | While AI can assist with image capture and post-processing, the task requires human judgment about what to document, from which angles, and under what lighting conditions to preserve evidentiary value. Current systems cannot autonomously decide framing and coverage for complex crime scenes without human oversight. |
| Task automatability | claude-sonnet-5 | 2/5 | Physically photographing and videoing a crime scene requires being present, moving through the space, and making real-time judgment calls about what to capture, which current AI cannot perform end-to-end.dea The task is fundamentally physical and manual, not something an AI system can execute autonomously. |
| Adoption barriers | claude-haiku-4-5-20251001 | 4/5 | Legal and regulatory frameworks require chain-of-custody documentation and often demand that a licensed forensic technician personally conduct and certify evidence photography. Liability asymmetry is severe: AI-captured evidence could be challenged in court, necessitating human sign-off. |
| Adoption barriers | claude-sonnet-5 | 4/5 | Evidence documentation is subject to strict chain-of-custody, forensic protocol, and legal admissibility requirements, typically demanding a trained, certified technician to perform and attest to the process. |
| Cost vs. human wage | claude-haiku-4-5-20251001 | 2/5 | High-quality forensic-grade cameras and video equipment remain expensive, and human expertise in scene framing and evidence prioritization cannot be easily replaced. Integration and liability oversight costs remain substantial relative to technician wages. |
| Cost vs. human wage | claude-sonnet-5 | 1/5 | There is no viable AI substitute performing the physical documentation, so cost comparison favors the human technician who must be physically present with equipment. |
| Technical feasibility today | claude-haiku-4-5-20251001 | 2/5 | Image capture hardware is mature, but no deployed product autonomously documents crime scenes with the precision, legal defensibility, and contextual awareness required for forensic purposes. Existing AI tools assist with processing but not with the intentional scene documentation itself. |
| Technical feasibility today | claude-sonnet-5 | 1/5 | No deployed product autonomously conducts crime scene photography or videography; this remains a human field task with cameras as tools, not AI-driven systems. |
Review forensic analysts' reports for technical merit.
19CI 13–25 · exposure 20 · augmentation 50 · importance 4.0/5 · click for rater detail
Review forensic analysts' reports for technical merit.
19| Dimension | Model | Rating | Rationale |
|---|---|---|---|
| Sector adoption velocity | claude-haiku-4-5-20251001 | 1/5 | Adoption is minimal in this highly regulated, small-population sector. Forensic labs operate under strict accreditation and legal compliance frameworks that have not yet normalized AI-assisted or AI-driven technical review, and organizational culture remains heavily human-expert-dependent. |
| Sector adoption velocity | claude-sonnet-5 | 2/5 | Forensic science labs are typically under-resourced government entities with slow technology adoption cycles, conservative practices, and strict accreditation requirements limiting AI integration. |
| Augmentation potential | claude-haiku-4-5-20251001 | 3/5 | AI can usefully assist by highlighting inconsistencies, flagging deviations from standard protocols, and summarizing report structure, which could help human reviewers work more efficiently. However, the core judgment about technical merit still rests with the human expert. |
| Augmentation potential | claude-sonnet-5 | 3/5 | AI can assist by flagging inconsistencies, checking documentation completeness, cross-referencing procedures, or summarizing reports, aiding the reviewer without replacing their judgment. |
| Task automatability | claude-haiku-4-5-20251001 | 2/5 | Reviewing forensic reports for technical merit requires nuanced judgment about methodology, statistical validity, and compliance with professional standards. While AI can extract and flag certain structural issues or inconsistencies, the critical assessment of whether analytical approaches are sound and appropriate for the evidence type remains heavily dependent on expert human judgment that current systems cannot reliably replicate end-to-end. |
| Task automatability | claude-sonnet-5 | 2/5 | Technical review requires expert judgment about methodology, chain of custody, and scientific validity that current AI cannot reliably assess end-to-end, though it could assist with checking formatting or completeness. |
| Adoption barriers | claude-haiku-4-5-20251001 | 4/5 | Strong barriers exist: forensic report review is often a legal requirement tied to expert certification and chain-of-custody standards. Courts and regulatory bodies typically require a credentialed human professional to vouch for the technical adequacy of forensic work, creating both liability and certification obstacles to full automation. |
| Adoption barriers | claude-sonnet-5 | 5/5 | Forensic report review typically requires certified/licensed technical reviewers per accreditation standards (e.g., ISO 17025, ASCLD/LAB), and legal admissibility in court demands qualified human sign-off. |
| Cost vs. human wage | claude-haiku-4-5-20251001 | 2/5 | Integration and oversight costs would be substantial given the high liability stakes; a trained forensic reviewer must still validate any AI-generated assessment, meaning cost approaches or exceeds that of human review alone rather than achieving significant savings. |
| Cost vs. human wage | claude-sonnet-5 | 2/5 | Even if AI could flag some issues, the liability and expertise requirements mean a qualified human reviewer is still needed, so cost savings are minimal or the AI adds cost as a supplementary tool rather than a replacement. |
| Technical feasibility today | claude-haiku-4-5-20251001 | 2/5 | No deployed product reliably performs independent technical review of forensic reports at production scale. AI systems can assist with document analysis and flagging of obvious errors, but validated production systems that perform independent technical merit assessment of forensic work do not yet exist in active use. |
| Technical feasibility today | claude-sonnet-5 | 1/5 | No deployed products perform substantive technical merit review of forensic reports in production; this remains a specialized expert function without commercial AI tools filling this role. |
Train new technicians or other personnel on forensic science techniques.
15CI 5–25 · exposure 13 · augmentation 63 · importance 4.3/5 · click for rater detail
Train new technicians or other personnel on forensic science techniques.
15| Dimension | Model | Rating | Rationale |
|---|---|---|---|
| Sector adoption velocity | claude-haiku-4-5-20251001 | 1/5 | Forensic science training remains heavily dependent on direct human mentorship and in-person lab work. Law enforcement and forensic labs are public-sector, regulation-heavy organizations that adopt new technologies slowly and have no incentive to remove human trainers from this mission-critical function. |
| Sector adoption velocity | claude-sonnet-5 | 2/5 | Forensic science labs are government/public-sector entities with slower technology adoption cycles and strict procedural requirements limiting AI integration. |
| Augmentation potential | claude-haiku-4-5-20251001 | 3/5 | AI can augment trainers by generating multimedia training materials, simulating case scenarios, or administering knowledge checks, thereby freeing trainers to focus on hands-on demonstration and individual assessment. However, the core mentoring and skill validation still requires human presence. |
| Augmentation potential | claude-sonnet-5 | 4/5 | AI can meaningfully augment training via generating quizzes, case studies, simulations, and reference materials, though it doesn't replace hands-on instruction. |
| Task automatability | claude-haiku-4-5-20251001 | 2/5 | Training others in forensic science techniques requires real-time adaptation, hands-on demonstration, assessment of individual learner progress, and mentoring—capabilities that current AI systems struggle with at scale. While AI can generate training materials or answer questions, conducting actual technical training with live feedback and adjustment falls well short of the ≥50% time-saving threshold. |
| Task automatability | claude-sonnet-5 | 1/5 | Training new forensic technicians requires hands-on demonstration of evidence handling, lab protocols, and physical technique that AI cannot deliver end-to-end today. |
| Adoption barriers | claude-haiku-4-5-20251001 | 4/5 | Training of forensic personnel involves legal and professional standards—new technicians must demonstrate competency to work in law enforcement and forensic labs. Licensing, certification, and liability requirements mean a qualified human instructor must supervise and sign off on training, creating a strong regulatory barrier. |
| Adoption barriers | claude-sonnet-5 | 4/5 | Forensic training often requires certified experts and adherence to accreditation standards (e.g., ASCLD/LAB) ensuring proper chain-of-custody and technique competency, creating strong institutional barriers. |
| Cost vs. human wage | claude-haiku-4-5-20251001 | 1/5 | An experienced forensic technician trainer's domain expertise, mentoring ability, and legal/liability responsibility cannot be replicated at lower cost by current AI systems. The human trainer remains essential, making AI replacement economically infeasible. |
| Cost vs. human wage | claude-sonnet-5 | 2/5 | AI-generated training content is cheap, but the actual training task still requires paid expert instructors and supervised practice, keeping overall costs comparable to human-led training. |
| Technical feasibility today | claude-haiku-4-5-20251001 | 2/5 | No deployed product reliably conducts live forensic training at production scale. AI systems can support training materials and simulations, but assessments of whether trainees can actually perform forensic techniques competently require human supervision and hands-on evaluation in real forensic contexts. |
| Technical feasibility today | claude-sonnet-5 | 1/5 | No deployed product performs live, hands-on forensic training; at best AI offers supplementary materials, not substitution for instructor-led training. |
Operate and maintain laboratory equipment and apparatus.
15CI 5–25 · exposure 13 · augmentation 25 · importance 4.1/5 · click for rater detail
Operate and maintain laboratory equipment and apparatus.
15| Dimension | Model | Rating | Rationale |
|---|---|---|---|
| Sector adoption velocity | claude-haiku-4-5-20251001 | 2/5 | Forensic labs are government agencies and specialized facilities with slow digitization and conservative adoption patterns due to regulatory constraints and legal liability concerns. Automation of core lab operations remains minimal despite general IT modernization, with most labs still relying on manual equipment operation. |
| Sector adoption velocity | claude-sonnet-5 | 1/5 | Forensic labs are a small, highly specialized, physically-oriented sector with minimal AI-driven automation of equipment operation to date. |
| Augmentation potential | claude-haiku-4-5-20251001 | 2/5 | AI offers limited augmentation for equipment operation and maintenance—perhaps basic monitoring alerts or predictive maintenance suggestions—but the core task of operating specialized forensic apparatus and troubleshooting equipment remains heavily dependent on human expertise and hands-on problem-solving. |
| Augmentation potential | claude-sonnet-5 | 2/5 | AI can help with equipment diagnostics, scheduling maintenance, or interpreting some output data, but offers little direct assistance with the physical operation and upkeep tasks themselves. |
| Task automatability | claude-haiku-4-5-20251001 | 2/5 | While some equipment operation steps (e.g., running pre-configured assays, recording readings) could be partially automated, forensic lab equipment requires calibration, troubleshooting, maintenance decisions, and quality assurance that demand human judgment and physical manipulation. Current AI systems cannot reliably handle the full range of maintenance tasks and equipment variability at 50% time savings. |
| Task automatability | claude-sonnet-5 | 1/5 | This is a hands-on physical task requiring manual manipulation, calibration, and maintenance of lab instruments (spectrometers, microscopes, chromatographs), which current AI cannot physically perform without robotics. |
| Adoption barriers | claude-haiku-4-5-20251001 | 4/5 | Forensic labs operate under strict regulatory oversight (ASCLD-LAB, FBI quality standards), chain-of-custody requirements, and legal admissibility standards that mandate documented human responsibility for equipment operation and maintenance. Many protocols legally require certified technicians to perform or directly oversee equipment operation. |
| Adoption barriers | claude-sonnet-5 | 4/5 | Chain-of-custody, accreditation standards, and forensic evidentiary requirements mean equipment operation must be performed and documented by certified technicians for legal admissibility. |
| Cost vs. human wage | claude-haiku-4-5-20251001 | 2/5 | Integration of robotic systems for lab equipment operation is capital-intensive and requires specialized customization for forensic workflows. The upfront and ongoing costs of equipment, maintenance, and supervision oversight typically exceed the cost of a trained forensic technician wage for this specific task. |
| Cost vs. human wage | claude-sonnet-5 | 1/5 | There is no viable AI substitute for physical equipment operation and maintenance, so AI cost is not comparable—human labor remains the only option. |
| Technical feasibility today | claude-haiku-4-5-20251001 | 2/5 | Limited deployed AI systems exist for autonomous lab equipment operation in forensic settings. While some robotic liquid handlers and automated analyzers exist in clinical labs, forensic labs involve diverse equipment, specialized protocols, and regulatory requirements where AI-driven autonomous operation is not yet reliably demonstrated in production. |
| Technical feasibility today | claude-sonnet-5 | 1/5 | No deployed products autonomously operate or maintain forensic lab equipment; this remains a manual technician function performed by trained humans. |
Prepare solutions, reagents, or sample formulations needed for laboratory work.
15CI 14–16 · exposure 16 · augmentation 25 · importance 3.4/5 · click for rater detail
Prepare solutions, reagents, or sample formulations needed for laboratory work.
15| Dimension | Model | Rating | Rationale |
|---|---|---|---|
| Sector adoption velocity | claude-haiku-4-5-20251001 | 1/5 | Forensic laboratories remain relatively low-digitization, physically-constrained environments with strong regulatory requirements for human oversight and documented procedures. Adoption of physical laboratory automation has been slow and limited to specialized large facilities. |
| Sector adoption velocity | claude-sonnet-5 | 2/5 | Forensic science labs are a niche, highly regulated, and physically-oriented sector with slow technology adoption cycles compared to information-based industries. |
| Augmentation potential | claude-haiku-4-5-20251001 | 2/5 | AI could assist by recommending precise formulations, reagent specifications, or flagging potential errors in preparation protocols before the technician begins work, but the core task of physically preparing and measuring solutions relies on the human's hands and judgment. |
| Augmentation potential | claude-sonnet-5 | 2/5 | AI can assist with calculating dilutions, tracking inventory, or generating protocols, but it offers limited direct assistance for the hands-on physical preparation task itself. |
| Task automatability | claude-haiku-4-5-20251001 | 2/5 | While AI could assist in calculating precise chemical formulations and reagent proportions, the task requires hands-on preparation, measurement, mixing, and often real-time sensory feedback (visual, tactile) that current AI systems cannot perform physically. Setup and supervision would be extensive and error-prone. |
| Task automatability | claude-sonnet-5 | 2/5 | Physical preparation of solutions and reagents requires manual manipulation, precise measurement, and handling of physical materials, which current AI cannot perform without robotic embodiment.cyber physical robotics is not mature for this precision lab work., |
| Adoption barriers | claude-haiku-4-5-20251001 | 4/5 | Forensic laboratory work is heavily regulated under chain-of-custody and quality assurance protocols; documented human preparation and verification of reagent formulations is often a legal or accreditation requirement. The liability and regulatory framework create substantial barriers to full automation. |
| Adoption barriers | claude-sonnet-5 | 4/5 | Forensic lab work is subject to strict chain-of-custody, quality control, and accreditation requirements (e.g., ISO 17025), often requiring certified technicians to prepare and verify reagents used in evidentiary analysis. |
| Cost vs. human wage | claude-haiku-4-5-20251001 | 1/5 | Current AI systems cannot physically perform this task, so comparative cost is not yet meaningful. Even with future automation, the capital cost of specialized laboratory robotics would far exceed the loaded wage of trained forensic technicians for the foreseeable future. |
| Cost vs. human wage | claude-sonnet-5 | 1/5 | Automating this requires specialized lab robotics and liquid-handling hardware that is expensive to acquire, integrate, and validate, making it costlier than a technician performing it directly today. |
| Technical feasibility today | claude-haiku-4-5-20251001 | 1/5 | No deployed AI product can autonomously prepare chemical solutions and reagents in a forensic laboratory setting. This is a physical task requiring laboratory equipment operation and precise handling that remains in the realm of robotic hardware research, not production deployment. |
| Technical feasibility today | claude-sonnet-5 | 1/5 | No deployed AI products autonomously prepare forensic laboratory reagents or sample formulations; this remains a manual, human-performed task with occasional automated liquid-handling robots in limited settings, not general AI. |
Examine physical evidence, such as hair, biological fluids, fiber, wood, or soil residues to obtain information about its source and composition.
13CI 0–25 · exposure 13 · augmentation 63 · importance 4.1/5 · click for rater detail
Examine physical evidence, such as hair, biological fluids, fiber, wood, or soil residues to obtain information about its source and composition.
13| Dimension | Model | Rating | Rationale |
|---|---|---|---|
| Sector adoption velocity | claude-haiku-4-5-20251001 | 2/5 | Forensic labs are cautious early adopters constrained by legal precedent and accreditation standards; AI integration is slow and focused on assisting analysts rather than replacing them, with limited public evidence of displacement. |
| Sector adoption velocity | claude-sonnet-5 | 1/5 | Forensic science labs are highly specialized, regulated, and slow to adopt automation for evidentiary physical analysis, with adoption limited to software aids rather than displacing the core physical task. |
| Augmentation potential | claude-haiku-4-5-20251001 | 4/5 | AI can meaningfully enhance technician productivity by automating image analysis, pattern matching in fiber and soil databases, and composition estimation from spectroscopic data, allowing the technician to focus on interpretation and legal documentation. |
| Augmentation potential | claude-sonnet-5 | 3/5 | AI can assist with pattern matching, database searches (e.g., DNA/fiber databases), and result interpretation support, but the physical examination itself is not augmented by current AI tools. |
| Task automatability | claude-haiku-4-5-20251001 | 2/5 | Current AI can assist with image analysis and classification of some materials (fibers, soil composition via spectroscopy), but the task requires hands-on sample handling, chain-of-custody management, and contextual judgment about evidence relevance that human technicians must perform in person. |
| Task automatability | claude-sonnet-5 | 1/5 | This requires hands-on physical manipulation of trace evidence using microscopy, chromatography, and lab instruments—current AI cannot perform physical examination or sample handling. |
| Adoption barriers | claude-haiku-4-5-20251001 | 4/5 | Chain-of-custody requirements, legal admissibility standards, and expert testimony obligations in court proceedings impose strict regulatory and liability controls; a licensed technician must document and often personally testify to evidence handling and analysis. |
| Adoption barriers | claude-sonnet-5 | 5/5 | Forensic evidence analysis requires certified technicians whose findings must meet legal/chain-of-custody and courtroom admissibility standards, with strict licensing and liability requirements. |
| Cost vs. human wage | claude-haiku-4-5-20251001 | 2/5 | Forensic technician labor is specialized and heavily regulated; AI tools for microscopy and chemical analysis require significant integration and oversight costs, and currently supplement rather than replace the technician's work, keeping total cost higher than the human alone. |
| Cost vs. human wage | claude-sonnet-5 | 1/5 | AI cannot replace the physical lab work and specialized instrumentation involved, so there is no viable cost comparison for full task substitution today. |
| Technical feasibility today | claude-haiku-4-5-20251001 | 2/5 | AI-assisted image analysis and material identification tools exist in research contexts, but deployed products for forensic evidence examination remain limited and require human validation; no fully autonomous system reliably handles the diversity of biological, textile, and soil samples at scale. |
| Technical feasibility today | claude-sonnet-5 | 1/5 | No deployed product autonomously conducts physical forensic evidence examination; this remains a manual laboratory science task performed by trained technicians. |
Examine and analyze blood stain patterns at crime scenes.
13CI 0–25 · exposure 13 · augmentation 38 · importance 3.9/5 · click for rater detail
Examine and analyze blood stain patterns at crime scenes.
13| Dimension | Model | Rating | Rationale |
|---|---|---|---|
| Sector adoption velocity | claude-haiku-4-5-20251001 | 2/5 | Forensic labs are traditional, resource-constrained organizations with strong professional norms requiring human expertise. Adoption of AI assistants is slow and remains largely in research partnerships rather than routine operational deployment. |
| Sector adoption velocity | claude-sonnet-5 | 1/5 | Forensic science is a highly specialized, low-digitization field with minimal AI agent deployment for physical scene analysis, reflecting slow adoption typical of hands-on public safety work. |
| Augmentation potential | claude-haiku-4-5-20251001 | 3/5 | AI can assist by automating preliminary image enhancement, pattern highlighting, and flagging anomalies, allowing the forensic technician to focus on interpretation and causal reasoning. Useful but not transformative, since expert judgment remains central. |
| Augmentation potential | claude-sonnet-5 | 2/5 | AI can assist with documentation, measurement calculations, or trajectory modeling software, but the core interpretive analysis and scene assessment remain largely human-driven with limited AI tool support currently in practice. |
| Task automatability | claude-haiku-4-5-20251001 | 2/5 | While AI can assist with image analysis and pattern recognition of bloodstains, the task requires expert spatial reasoning, contextual judgment about crime scene dynamics, and often testimony in court. Current systems cannot reliably replace the full investigative and interpretive work, though they can flag features for human review. |
| Task automatability | claude-sonnet-5 | 1/5 | Blood stain pattern analysis requires physical presence at the scene, spatial judgment, and expert interpretation of 3D geometry that current AI cannot perform end-to-end without a human examiner directing the process. |
| Adoption barriers | claude-haiku-4-5-20251001 | 4/5 | Legal and professional barriers are substantial: courts require expert testimony from qualified, licensed forensic technicians; results must often be personally certified and defended by a human expert. Liability for incorrect analysis and the legal admissibility standard place hard constraints on automation. |
| Adoption barriers | claude-sonnet-5 | 5/5 | Forensic evidence used in court requires certified, qualified human experts whose findings are subject to legal admissibility standards, chain-of-custody, and expert testimony requirements. |
| Cost vs. human wage | claude-haiku-4-5-20251001 | 2/5 | AI-assisted analysis tools exist but require significant setup, calibration, and human oversight. The cost of deployment plus forensic scientist review time remains comparable to or exceeds the cost of having a trained technician perform the analysis directly. |
| Cost vs. human wage | claude-sonnet-5 | 1/5 | There is no viable AI substitute performing this task, so any AI cost is not comparable to a human forensic technician's wage for equivalent output. |
| Technical feasibility today | claude-haiku-4-5-20251001 | 2/5 | Some tools exist for bloodstain image processing and pattern matching, but no deployed system reliably performs the full analytical task—determination of impact angle, blood origin, sequence of events—at production quality. Most applications remain research or proof-of-concept stages. |
| Technical feasibility today | claude-sonnet-5 | 1/5 | No deployed product autonomously conducts crime scene bloodstain pattern analysis; this remains a specialized forensic skill performed manually by trained technicians. |
Examine footwear, tire tracks, or other types of impressions.
12CI 4–20 · exposure 13 · augmentation 50 · importance 3.3/5 · click for rater detail
Examine footwear, tire tracks, or other types of impressions.
12| Dimension | Model | Rating | Rationale |
|---|---|---|---|
| Sector adoption velocity | claude-haiku-4-5-20251001 | 2/5 | Forensic labs are traditionally conservative, under-resourced, and subject to strict procedural compliance. While some labs pilot AI visualization tools, production deployment of AI-driven impression analysis remains rare; adoption is cautious and typically limited to research or ancillary screening roles. |
| Sector adoption velocity | claude-sonnet-5 | 1/5 | Forensic science labs are a small, specialized, slow-to-digitize sector with strict procedural and legal constraints, resulting in minimal AI deployment for this task. |
| Augmentation potential | claude-haiku-4-5-20251001 | 3/5 | AI can meaningfully assist by enhancing image clarity, highlighting candidate features, or flagging potential matches for human expert review, thereby accelerating the technician's workflow. However, the augmentation is bounded by the need for expert final judgment on each comparison. |
| Augmentation potential | claude-sonnet-5 | 3/5 | AI-assisted image comparison and pattern-matching tools can help technicians narrow down candidate matches and speed preliminary analysis, though final judgment remains human-driven. |
| Task automatability | claude-haiku-4-5-20251001 | 2/5 | While AI can assist with image analysis and pattern recognition on clear, well-documented impression samples, the task requires expert judgment about 3D geometry, material properties, wear patterns, and chain-of-custody verification that current systems cannot reliably execute end-to-end. The variability of impression conditions, lighting, and substrate makes consistent automation below the 50% threshold. |
| Task automatability | claude-sonnet-5 | 1/5 | Physical examination of impressions requires on-site collection, casting, and forensic comparison expertise that current AI cannot perform end-to-end; image analysis alone doesn't cover the full task workflow.rulesof evidence handling.RESULT |
| Adoption barriers | claude-haiku-4-5-20251001 | 5/5 | Forensic evidence analysis is subject to legal admissibility standards (Daubert, evidence rules) that require human expert certification and courtroom testimony. Chain-of-custody, legal liability for misidentification, and regulatory requirements mean a licensed forensic technician must authenticate and vouch for findings regardless of AI preprocessing. |
| Adoption barriers | claude-sonnet-5 | 5/5 | Forensic examination results are used as legal evidence and typically require certified/qualified examiners to testify and sign off, creating strong legal and licensing barriers to automation. |
| Cost vs. human wage | claude-haiku-4-5-20251001 | 2/5 | Current AI-assisted impression analysis tools require significant integration, domain-specific training data, and expert technician oversight to validate outputs. The per-case cost remains comparable to or exceeds human examination once all infrastructure and quality control are factored in. |
| Cost vs. human wage | claude-sonnet-5 | 2/5 | AI tools could someday cheaply assist in image matching, but the physical evidence collection, chain-of-custody, and expert testimony components still require costly trained human labor. |
| Technical feasibility today | claude-haiku-4-5-20251001 | 2/5 | Some research tools and limited commercial products exist for impression comparison and enhancement, but they function as assistive visualization aids rather than reliable decision systems. Deployed forensic labs use these outputs with mandatory human expert review; no product autonomously determines matches or exclusions in production workflows. |
| Technical feasibility today | claude-sonnet-5 | 1/5 | No deployed AI product autonomously examines footwear or tire impressions in forensic casework today; comparison software exists only as decision-support, not autonomous examination. |
Reconstruct crime scenes to determine relationships among pieces of evidence.
11CI 3–20 · exposure 13 · augmentation 63 · importance 4.1/5 · click for rater detail
Reconstruct crime scenes to determine relationships among pieces of evidence.
11| Dimension | Model | Rating | Rationale |
|---|---|---|---|
| Sector adoption velocity | claude-haiku-4-5-20251001 | 2/5 | Forensic labs are typically government agencies or tightly regulated private firms with slow digital transformation and high compliance overhead. While some labs use 3D imaging and data management tools, adoption of AI-driven automation of core reconstruction work remains minimal and hesitant—laggard sectors relative to information and finance. |
| Sector adoption velocity | claude-sonnet-5 | 2/5 | Forensic science is a specialized, government-adjacent field with slow technology adoption cycles, cautious validation requirements, and limited deployment of AI in operational crime scene work. |
| Augmentation potential | claude-haiku-4-5-20251001 | 4/5 | AI tools already augment human forensic technicians meaningfully: 3D photogrammetry software, pattern recognition on evidence images, spatial visualization aids, and data management systems increase productivity and accuracy in documentation and preliminary analysis. These assistive tools are deployed today and meaningfully elevate technician capability while leaving judgment and legal authority with the human expert. |
| Augmentation potential | claude-sonnet-5 | 3/5 | AI can assist with 3D scene modeling, evidence cataloging, pattern matching, and simulation to support technicians' reconstruction hypotheses, offering meaningful but partial productivity gains. |
| Task automatability | claude-haiku-4-5-20251001 | 2/5 | Crime scene reconstruction requires spatial reasoning, contextual judgment, and integration of multiple evidence types—capabilities that current AI cannot reliably perform end-to-end. While AI can assist with image analysis and 3D modeling of physical spaces, the interpretive work of determining causal relationships between evidence pieces and reconstructing sequences of events demands human expertise and legal defensibility that AI cannot yet provide at production quality. |
| Task automatability | claude-sonnet-5 | 1/5 | Crime scene reconstruction requires physical presence, spatial reasoning across ambiguous physical evidence, and expert judgment integrating forensic disciplines; current AI cannot perform this end-to-end.6ecause it depends heavily on physical inspection and multidisciplinary inference no off-the-shelf system approaches the 50% time-saving bar.rating 1 |
| Adoption barriers | claude-haiku-4-5-20251001 | 5/5 | Forensic evidence reconstruction is heavily regulated and legally scrutinized; findings must be defensible in court and typically require testimony by a certified human expert. Liability for errors is high, chain-of-custody requirements are strict, and most jurisdictions require a licensed forensic technician to prepare and sign off on reconstructions. These are hard legal barriers to automation. |
| Adoption barriers | claude-sonnet-5 | 5/5 | Crime scene reconstruction findings are used as legal evidence, requiring certified forensic technicians to testify and be accountable under chain-of-custody and evidentiary rules, creating hard legal and licensing barriers. |
| Cost vs. human wage | claude-haiku-4-5-20251001 | 2/5 | AI assistance tools (3D scanning, image analysis software) require substantial infrastructure, training data, and oversight. The cost of deploying and validating such systems—plus the human expert time still required for interpretation and legal defensibility—remains comparable to or higher than the cost of human technicians working alone. |
| Cost vs. human wage | claude-sonnet-5 | 1/5 | There is no viable AI substitute performing this task, so cost comparison favors the human forensic technician entirely; any AI-assisted tools add cost without replacing the core analytical work. |
| Technical feasibility today | claude-haiku-4-5-20251001 | 2/5 | No deployed AI product today reliably performs full crime scene reconstruction. Computer vision and 3D imaging tools exist as narrow aids, but no system combines evidence analysis, spatial reconstruction, and causal reasoning in a way that meets forensic or legal standards. This remains primarily a human expert domain with only emerging tool support. |
| Technical feasibility today | claude-sonnet-5 | 1/5 | No deployed product autonomously reconstructs crime scenes; this remains a research-stage application at best, with any AI tools limited to narrow visualization or simulation aids rather than actual reconstruction judgment. |
Analyze gunshot residue and bullet paths to determine how shootings occurred.
10CI 0–20 · exposure 13 · augmentation 50 · importance 3.9/5 · click for rater detail
Analyze gunshot residue and bullet paths to determine how shootings occurred.
10| Dimension | Model | Rating | Rationale |
|---|---|---|---|
| Sector adoption velocity | claude-haiku-4-5-20251001 | 2/5 | Forensic science agencies and law enforcement have been slow to adopt autonomous AI for core analysis tasks, with most AI use limited to assistive image processing and simulation aids. Integration remains fragmented across agencies, budgets are constrained, and courts' preference for human expert testimony slows organizational incentive to automate. |
| Sector adoption velocity | claude-sonnet-5 | 1/5 | Forensic science labs are slow-adopting, government-funded, physically grounded environments with minimal AI agent deployment for evidentiary analysis. |
| Augmentation potential | claude-haiku-4-5-20251001 | 4/5 | AI significantly augments forensic technicians through ballistic trajectory simulation, gunshot residue particle detection in images, and 3D crime scene reconstruction visualization. These tools measurably improve technician productivity and confidence while the human expert retains control, interpretation, and legal accountability for conclusions. |
| Augmentation potential | claude-sonnet-5 | 2/5 | AI can assist with some data analysis, pattern matching in trajectory calculations, or report drafting, but core residue analysis and scene interpretation remain manual and instrument-driven. |
| Task automatability | claude-haiku-4-5-20251001 | 2/5 | While AI can assist with pattern recognition in gunshot residue imagery and ballistic trajectory modeling, the task requires expert judgment integrating multiple evidence types, crime scene context, and interpretive decisions that current AI cannot reliably automate end-to-end. A forensic technician must synthesize physical evidence, validate assumptions, and provide testimony-ready conclusions—functions requiring human expertise and accountability. |
| Task automatability | claude-sonnet-5 | 1/5 | This requires physical evidence collection, chemical analysis (SEM/EDX), and complex spatial/ballistic reconstruction at crime scenes that current AI cannot perform end-to-end.' |
| Adoption barriers | claude-haiku-4-5-20251001 | 5/5 | Forensic analysis for criminal proceedings faces hard regulatory and legal barriers: testimony must typically be delivered by a licensed/certified forensic examiner; chain-of-custody and evidence handling are legally mandated; and courts require human expert accountability. AI outputs cannot independently serve as forensic evidence without expert human endorsement and signature. |
| Adoption barriers | claude-sonnet-5 | 5/5 | Forensic evidence used in criminal prosecutions requires certified, legally accountable experts who can testify in court and maintain chain of custody, creating hard legal and licensing barriers. |
| Cost vs. human wage | claude-haiku-4-5-20251001 | 2/5 | Current AI tools for ballistic analysis and residue imaging are expensive specialized software requiring trained operators and substantial setup. When factored against the loaded cost of a forensic technician, the all-in AI cost per case analysis is comparable or higher, especially given the need for human validation and expert sign-off. |
| Cost vs. human wage | claude-sonnet-5 | 1/5 | AI has no capability to substitute for the physical inspection, sampling, and instrument-based chemical analysis involved, so cost comparison favors human technicians entirely. |
| Technical feasibility today | claude-haiku-4-5-20251001 | 2/5 | AI tools exist for image analysis and ballistic simulation (e.g., trajectory software), but no end-to-end deployed system reliably performs this entire forensic analysis independently. Products are narrow, require expert oversight, and are not used in production to replace human technician judgment on critical case conclusions. |
| Technical feasibility today | claude-sonnet-5 | 1/5 | No deployed AI product performs gunshot residue analysis or bullet trajectory reconstruction; this remains a specialized forensic lab and field task done by trained technicians. |
Compare objects, such as tools, with impression marks to determine whether a specific object is responsible for a specific mark.
9CI 0–18 · exposure 13 · augmentation 38 · importance 3.7/5 · click for rater detail
Compare objects, such as tools, with impression marks to determine whether a specific object is responsible for a specific mark.
9| Dimension | Model | Rating | Rationale |
|---|---|---|---|
| Sector adoption velocity | claude-haiku-4-5-20251001 | 1/5 | Forensic labs are typically government or institutional operations with slow digital transformation, conservative practices, and resistance to replacing expert judgment in case-critical tasks. Adoption of AI-assisted (not replacing) tools is emerging slowly in some labs but remains far behind information and professional services sectors. |
| Sector adoption velocity | claude-sonnet-5 | 1/5 | Forensic science labs are slow-adopting, tightly regulated, and physically evidence-based, with minimal AI agent deployment in actual casework today. |
| Augmentation potential | claude-haiku-4-5-20251001 | 3/5 | AI image alignment and feature highlighting can assist examiners by automating initial comparison steps and flagging candidate matches, reducing manual visual strain and speeding preliminary screening. However, the core judgment—whether a match is conclusive—remains with the human expert. |
| Augmentation potential | claude-sonnet-5 | 2/5 | Digital imaging and measurement software can assist in documenting and comparing marks, offering some efficiency, but the core judgment task still relies heavily on the human examiner's expertise. |
| Task automatability | claude-haiku-4-5-20251001 | 2/5 | Comparing objects with impression marks requires nuanced visual and spatial analysis, but current computer vision can detect and align physical features. However, the task demands judgment about match quality, exclusion of false positives, and forensic certainty that goes beyond pattern matching—human expertise in interpreting ambiguity and weighing evidence remains essential for reliable results. |
| Task automatability | claude-sonnet-5 | 1/5 | Physical tool-mark comparison requires microscopic examination and expert pattern judgment on physical evidence; no off-the-shelf AI system can perform this end-to-end today with equal quality time savings.' |
| Adoption barriers | claude-haiku-4-5-20251001 | 5/5 | Forensic evidence admissibility is tightly regulated; courts require testimony from qualified examiners, and many jurisdictions demand human expert certification and courtroom cross-examination. Liability and error costs in criminal/civil cases are asymmetrically high, creating a hard legal barrier to unsupervised AI decision-making. |
| Adoption barriers | claude-sonnet-5 | 5/5 | Forensic conclusions used in criminal justice require certified, qualified examiners who can testify and be cross-examined; legal and evidentiary standards mandate human expert sign-off. |
| Cost vs. human wage | claude-haiku-4-5-20251001 | 2/5 | Forensic impression comparison requires specialized imaging hardware, expert validation, and legal oversight. AI software might reduce some comparison time, but the full workflow—including equipment, training, legal review, and expert sign-off—remains expensive relative to incremental automation cost savings. |
| Cost vs. human wage | claude-sonnet-5 | 1/5 | Any AI-assisted imaging tool would still require specialized microscopy equipment, human verification, and legal defensibility, making it more costly than trained technician labor for now. |
| Technical feasibility today | claude-haiku-4-5-20251001 | 2/5 | While AI-powered image comparison and alignment tools exist in research and some forensic labs, deployed products lack the reliability and courtroom acceptance required for definitive forensic conclusions. Most operational comparisons still rely on human examiners using traditional or semi-automated tools, not end-to-end AI systems. |
| Technical feasibility today | claude-sonnet-5 | 1/5 | This remains research-stage; while some image-based matching algorithms exist in academic studies, no deployed forensic product reliably performs courtroom-grade tool-mark identification in production labs. |
Collect impressions of dust from surfaces to obtain and identify fingerprints.
3CI 0–5 · exposure 0 · augmentation 25 · importance 4.1/5 · click for rater detail
Collect impressions of dust from surfaces to obtain and identify fingerprints.
3| Dimension | Model | Rating | Rationale |
|---|---|---|---|
| Sector adoption velocity | claude-haiku-4-5-20251001 | 1/5 | Forensic science operates in a highly regulated, physical, and human-supervised environment. Adoption of automation for evidence collection itself is not occurring in production; the sector is laggard on this particular task. |
| Sector adoption velocity | claude-sonnet-5 | 1/5 | Forensic field evidence collection is a highly manual, physical, small-scale domain with minimal AI/robotic deployment or investment reported for this specific step. |
| Augmentation potential | claude-haiku-4-5-20251001 | 2/5 | AI tools can assist in *analyzing* collected fingerprints (comparison, database matching), but during the collection phase itself—dusting, lifting, handling—AI offers minimal assistance to the technician beyond perhaps image documentation. |
| Augmentation potential | claude-sonnet-5 | 2/5 | AI can assist afterward with fingerprint matching/analysis via AFIS-type systems, but it offers little help with the physical dusting and lifting process itself. |
| Task automatability | claude-haiku-4-5-20251001 | 1/5 | This task requires physical manipulation in a crime scene environment, precision tactile work (dusting surfaces, lifting impressions), and real-time judgment about surface conditions. Current AI systems cannot physically interact with environments or perform the delicate, context-sensitive handling that fingerprint collection demands. |
| Task automatability | claude-sonnet-5 | 1/5 | This is a hands-on physical evidence collection task requiring dusting powders, lifting tape, and careful physical technique at a crime scene; no AI system can perform the physical manipulation involved. |
| Adoption barriers | claude-haiku-4-5-20251001 | 5/5 | Chain of custody and evidentiary integrity requirements legally mandate that a qualified, credentialed forensic technician collect, document, and sign off on physical evidence. Regulatory and liability frameworks are explicit and hard. |
| Adoption barriers | claude-sonnet-5 | 4/5 | Chain-of-custody, forensic certification standards, and legal admissibility requirements mean specially trained/certified personnel must perform and document evidence collection for court purposes. |
| Cost vs. human wage | claude-haiku-4-5-20251001 | 1/5 | The physical automation hardware (robotic arms, specialized dusting mechanisms) would be expensive to deploy at crime scenes, and integration complexity is high relative to the cost of a forensic technician performing the collection. |
| Cost vs. human wage | claude-sonnet-5 | 1/5 | There is no AI substitute for physical dust collection, so any comparison would require robotic hardware that does not exist for this task, making AI far more expensive or simply unavailable. |
| Technical feasibility today | claude-haiku-4-5-20251001 | 1/5 | No deployed AI products perform physical fingerprint collection. While AI can analyze already-collected fingerprints, the collection itself—the core of this task—remains entirely manual and requires a trained human technician in situ. |
| Technical feasibility today | claude-sonnet-5 | 1/5 | No deployed product performs physical fingerprint dusting or lifting; this remains a manual forensic technician skill with robots or AI not used for this specific physical collection step. |
Examine firearms to determine mechanical condition and legal status, performing restoration work on damaged firearms to obtain information, such as serial numbers.
3CI 0–5 · exposure 0 · augmentation 25 · importance 3.6/5 · click for rater detail
Examine firearms to determine mechanical condition and legal status, performing restoration work on damaged firearms to obtain information, such as serial numbers.
3| Dimension | Model | Rating | Rationale |
|---|---|---|---|
| Sector adoption velocity | claude-haiku-4-5-20251001 | 1/5 | Forensic science is a regulated, slow-moving sector with entrenched practices, legal requirements, and minimal AI adoption in core evidence-handling tasks. Adoption of automation in firearms forensics is negligible. |
| Sector adoption velocity | claude-sonnet-5 | 1/5 | Forensic laboratory physical evidence work is a low-digitization, highly manual field with essentially no AI/robotic adoption for hands-on firearm restoration tasks. |
| Augmentation potential | claude-haiku-4-5-20251001 | 2/5 | AI could provide limited assistance in image capture and documentation of firearm condition, but cannot augment the core work of mechanical restoration, hands-on assessment, or serial number recovery—the heart of this task. |
| Augmentation potential | claude-sonnet-5 | 2/5 | AI may assist with database lookups (e.g., matching recovered serial numbers to registries) or documentation, but offers minimal help with the core physical restoration and mechanical assessment work. |
| Task automatability | claude-haiku-4-5-20251001 | 1/5 | This task requires hands-on examination and physical manipulation of firearms, restoration work on damaged components, and expert judgment about mechanical condition—all fundamentally physical and tactile operations that current AI systems cannot perform end-to-end. While image analysis of firearms exists, the restoration and mechanical assessment components remain entirely human-dependent. |
| Task automatability | claude-sonnet-5 | 1/5 | This requires hands-on physical manipulation of firearms, mechanical disassembly, and restoration techniques (e.g., acid etching for serial numbers) that AI cannot perform without robotic embodiment far beyond current capability. |
| Adoption barriers | claude-haiku-4-5-20251001 | 5/5 | Firearms examination and restoration are restricted to licensed forensic professionals in law enforcement and certified forensic laboratories, with strict chain-of-custody and legal requirements. Regulatory and licensing barriers are hard—only authorized personnel may legally handle and examine firearms evidence. |
| Adoption barriers | claude-sonnet-5 | 4/5 | Chain-of-custody, evidentiary integrity, and legal admissibility requirements mean this work must be performed and attested to by qualified, often certified forensic personnel, creating strong procedural and legal barriers. |
| Cost vs. human wage | claude-haiku-4-5-20251001 | 1/5 | AI has no viable alternative deployment for this task; the cost comparison is moot. A trained forensic technician's labor remains the only method available, making any attempted AI solution more expensive than the status quo. |
| Cost vs. human wage | claude-sonnet-5 | 1/5 | There is no AI substitute performing this physical task, so cost comparison favors the human technician entirely since no viable AI alternative exists. |
| Technical feasibility today | claude-haiku-4-5-20251001 | 1/5 | No deployed AI product reliably performs firearm restoration, mechanical assessment, or serial number recovery. These tasks require specialized physical intervention and expert forensic knowledge that current AI systems, even with robotics integration, do not demonstrate in production environments. |
| Technical feasibility today | claude-sonnet-5 | 1/5 | No deployed AI product performs physical firearm restoration or mechanical examination; this remains entirely a manual forensic lab task performed by trained technicians. |
Collect evidence from crime scenes, storing it in conditions that preserve its integrity.
0CI 0–0 · exposure 0 · augmentation 38 · importance 4.9/5 · click for rater detail
Collect evidence from crime scenes, storing it in conditions that preserve its integrity.
0| Dimension | Model | Rating | Rationale |
|---|---|---|---|
| Sector adoption velocity | claude-haiku-4-5-20251001 | 1/5 | Law enforcement and forensic laboratories are slow to adopt automation in evidence collection due to strict regulatory oversight, liability concerns, and the legal requirement for human accountability in the chain-of-custody. |
| Sector adoption velocity | claude-sonnet-5 | 1/5 | Forensic fieldwork is a highly physical, low-digitization task with essentially no AI/robotic adoption for on-scene evidence handling. |
| Augmentation potential | claude-haiku-4-5-20251001 | 3/5 | AI can assist forensic technicians through evidence documentation, automated photography, data logging, and analytical support post-collection, but the core collection task remains human-dependent due to legal requirements and environmental complexity. |
| Augmentation potential | claude-sonnet-5 | 2/5 | AI can assist with documentation, logging, or scene photo analysis afterward, but offers minimal help during the physical act of evidence collection and preservation itself. |
| Task automatability | claude-haiku-4-5-20251001 | 1/5 | Forensic evidence collection requires physical manipulation of objects in complex, contaminated environments, spatial reasoning, and real-time decision-making about chain-of-custody that current AI cannot perform end-to-end. The task fundamentally depends on human judgment about what constitutes relevant evidence and proper handling protocols. |
| Task automatability | claude-sonnet-5 | 1/5 | Physical collection, chain-of-custody handling, and preservation of crime scene evidence require in-person manipulation of physical objects in unpredictable environments, which current AI cannot perform. |
| Adoption barriers | claude-haiku-4-5-20251001 | 5/5 | Strict legal and regulatory requirements mandate that evidence collection be performed by or under direct supervision of certified forensic technicians to preserve chain-of-custody; automation would violate established evidence handling laws and courtroom admissibility standards. |
| Adoption barriers | claude-sonnet-5 | 5/5 | Legal chain-of-custody requirements, evidentiary admissibility rules, and certification/training mandates require a qualified human technician to physically collect and document evidence. |
| Cost vs. human wage | claude-haiku-4-5-20251001 | 1/5 | The cost of deploying robotic systems capable of safe evidence handling, plus oversight and verification, far exceeds the loaded wage of a forensic technician, especially given the mission-critical nature of the work. |
| Cost vs. human wage | claude-sonnet-5 | 1/5 | There is no viable AI substitute performing this physical task, so cost comparison favors human technicians by default. |
| Technical feasibility today | claude-haiku-4-5-20251001 | 1/5 | No deployed products reliably perform autonomous forensic evidence collection at crime scenes. While imaging and documentation AI exist, no system can independently collect physical evidence while maintaining legal chain-of-custody integrity in production. |
| Technical feasibility today | claude-sonnet-5 | 1/5 | No deployed AI or robotic products perform physical crime scene evidence collection; this remains entirely a research-stage or non-existent capability. |
Testify in court about investigative or analytical methods or findings.
0CI 0–0 · exposure 0 · augmentation 38 · importance 4.5/5 · click for rater detail
Testify in court about investigative or analytical methods or findings.
0| Dimension | Model | Rating | Rationale |
|---|---|---|---|
| Sector adoption velocity | claude-haiku-4-5-20251001 | 1/5 | There is no adoption of AI to perform courtroom testimony because the task is legally non-automatable. Any adoption discussion centers on AI-assisted report drafting or summary for human testimony, not replacement of the testimony act itself. |
| Sector adoption velocity | claude-sonnet-5 | 1/5 | Legal and forensic testimony contexts are highly regulated and slow to change; there is no meaningful movement toward AI performing testimony itself. |
| Augmentation potential | claude-haiku-4-5-20251001 | 2/5 | AI can assist by preparing summaries, organizing evidence, or drafting expert report sections that the technician then presents, but the core testimony act—oral, adversarial, accountable—remains entirely human-driven with minimal direct augmentation of the testimony itself. |
| Augmentation potential | claude-sonnet-5 | 3/5 | AI can help technicians prepare testimony by organizing findings, drafting reports, or simulating cross-examination questions, but it doesn't change the courtroom act itself. |
| Task automatability | claude-haiku-4-5-20251001 | 1/5 | Courtroom testimony requires human credibility, cross-examination response, legal standing, and real-time judgment under adversarial conditions. AI cannot take an oath, be cross-examined, or serve as a legal witness—this is a fundamental legal and human requirement that no AI system can fulfill. |
| Task automatability | claude-sonnet-5 | 1/5 | Testifying in court is a live, sworn, in-person act by a qualified human expert; AI cannot legally or practically perform this end-to-end today. |
| Adoption barriers | claude-haiku-4-5-20251001 | 5/5 | Courtroom testimony is protected by hard legal barriers: only a human under oath can serve as a witness, and testimony requires cross-examination and legal accountability that AI cannot satisfy. Jurisdiction, evidence rules, and the defendant's right to confront witnesses all require a human witness. |
| Adoption barriers | claude-sonnet-5 | 5/5 | Courts require a qualified, credentialed human expert to testify under oath and be subject to cross-examination; this is a hard legal/licensing barrier. |
| Cost vs. human wage | claude-haiku-4-5-20251001 | 1/5 | Human forensic technicians must testify in person (or via authorized remote means) as legal witnesses; there is no meaningful AI cost comparison because AI cannot legally substitute for this function. The cost remains entirely human. |
| Cost vs. human wage | claude-sonnet-5 | 1/5 | There is no viable AI substitute cost to compare; the human expert witness is the only legally acceptable option, making AI effectively infinitely costlier in practice. |
| Technical feasibility today | claude-haiku-4-5-20251001 | 1/5 | No deployed product can perform courtroom testimony as an autonomous entity. While AI can draft written reports or synthesize findings for human testimony, the act of testifying as a witness is legally reserved to humans and does not exist as an automated AI service in any jurisdiction. |
| Technical feasibility today | claude-sonnet-5 | 1/5 | No deployed product testifies in court on behalf of a forensic expert; this remains purely hypothetical/research-stage at best. |
Visit morgues, examine scenes of crimes, or contact other sources to obtain evidence or information to be used in investigations.
0CI 0–0 · exposure 0 · augmentation 25 · importance 4.3/5 · click for rater detail
Visit morgues, examine scenes of crimes, or contact other sources to obtain evidence or information to be used in investigations.
0| Dimension | Model | Rating | Rationale |
|---|---|---|---|
| Sector adoption velocity | claude-haiku-4-5-20251001 | 1/5 | Adoption velocity is minimal because the fundamental requirement—physical, autonomous presence at crime scenes and morgues—is not automatable. Forensic agencies have strong institutional and legal incentives to maintain human investigation practices rather than pursue AI substitution. |
| Sector adoption velocity | claude-sonnet-5 | 1/5 | Forensic fieldwork is a physical, highly regulated, low-digitization task with essentially no AI/robotic adoption for on-site evidence gathering. |
| Augmentation potential | claude-haiku-4-5-20251001 | 2/5 | AI can assist tangentially (e.g., analyzing crime scene photos post-collection, organizing evidence databases, suggesting investigation leads), but cannot augment the core task of physically visiting locations and gathering evidence in real time. Assistance is limited to downstream analysis rather than the investigation itself. |
| Augmentation potential | claude-sonnet-5 | 2/5 | AI can assist with documentation, note transcription, or cross-referencing information sources afterward, but offers little help during the physical scene investigation itself. |
| Task automatability | claude-haiku-4-5-20251001 | 1/5 | This task involves physical presence at crime scenes and morgues, direct evidence handling, and real-time environmental assessment that requires human judgment and legal chain-of-custody protocols. Current AI systems cannot autonomously visit locations, examine physical evidence, or conduct the investigative interviews embedded in this work. |
| Task automatability | claude-sonnet-5 | 1/5 | This requires physical presence at crime scenes and morgues, physical evidence handling, and situational judgment that current AI cannot perform end-to-end. |
| Adoption barriers | claude-haiku-4-5-20251001 | 5/5 | This task is heavily protected by legal requirements (chain of custody, evidence handling protocols, witness credibility), professional licensing of forensic technicians, and statutory mandates that law enforcement personnel must physically investigate crime scenes. Court admissibility of evidence depends on documented human oversight. |
| Adoption barriers | claude-sonnet-5 | 5/5 | Chain-of-custody, legal admissibility, licensing, and forensic certification requirements mean only authorized trained personnel can legally collect and handle evidence. |
| Cost vs. human wage | claude-haiku-4-5-20251001 | 1/5 | There is no AI alternative to this task; human forensic technicians must physically attend scenes and morgues. Cost comparison is not meaningful because substitution is impossible, making AI more expensive than zero relevant deployment. |
| Cost vs. human wage | claude-sonnet-5 | 1/5 | AI has no capability to substitute for physical evidence collection, so there is no cost comparison favoring AI; humans remain the only viable option. |
| Technical feasibility today | claude-haiku-4-5-20251001 | 1/5 | No deployed AI system can independently visit crime scenes, conduct morgue examinations, or gather evidence through in-person investigation. While AI can assist with post-collection evidence analysis, the core task—obtaining evidence through site visits and source contact—remains entirely human-dependent. |
| Technical feasibility today | claude-sonnet-5 | 1/5 | No deployed product visits crime scenes or morgues to collect physical evidence; this remains entirely a human, on-site task. |
Confer with ballistics, fingerprinting, handwriting, documents, electronics, medical, chemical, or metallurgical experts concerning evidence and its interpretation.
0CI 0–0 · exposure 0 · augmentation 38 · importance 3.8/5 · click for rater detail
Confer with ballistics, fingerprinting, handwriting, documents, electronics, medical, chemical, or metallurgical experts concerning evidence and its interpretation.
0| Dimension | Model | Rating | Rationale |
|---|---|---|---|
| Sector adoption velocity | claude-haiku-4-5-20251001 | 1/5 | Forensic science operates in a highly regulated, liability-conscious environment (law enforcement, courts) where automation of expert consultation faces strong institutional and legal resistance. Adoption of AI for this task is negligible in production systems. |
| Sector adoption velocity | claude-sonnet-5 | 1/5 | Forensic science is a specialized, tightly regulated, low-digitization field where AI adoption for expert consultation processes is minimal and largely absent from public sector crime labs. |
| Augmentation potential | claude-haiku-4-5-20251001 | 2/5 | AI might assist with organizing evidence summaries or flagging prior case comparisons, but the core interpretive and consultative aspects of this task depend on expert human judgment and professional accountability that AI cannot meaningfully augment without substantial human rework. |
| Augmentation potential | claude-sonnet-5 | 3/5 | AI tools can help summarize case files, flag discrepancies, or assist in preparing questions for expert consultations, offering moderate support without replacing the human interaction. |
| Task automatability | claude-haiku-4-5-20251001 | 1/5 | This task requires substantive expert judgment, professional reasoning about complex evidence interpretation, and collegial consultation between trained specialists. Current AI cannot reliably perform the core function of conferring with experts and synthesizing their interpretations into coherent forensic conclusions. |
| Task automatability | claude-sonnet-5 | 1/5 | This is an interdisciplinary consultative task requiring real-time professional judgment, trust-building, and synthesis of expert opinions across specialized forensic domains, which current AI cannot replicate end-to-end. |
| Adoption barriers | claude-haiku-4-5-20251001 | 5/5 | Forensic evidence interpretation carries high legal liability; expert testimony and chain-of-custody documentation are legally mandated to come from qualified human experts. Regulatory and evidentiary standards require human professionals to sign off on and testify to forensic findings. |
| Adoption barriers | claude-sonnet-5 | 5/5 | Forensic evidence interpretation often has legal chain-of-custody and expert-testimony requirements, meaning credentialed human experts must be involved and accountable in court-admissible determinations. |
| Cost vs. human wage | claude-haiku-4-5-20251001 | 1/5 | Current AI systems would require significant human oversight, validation, and expert input to be trustworthy in a forensic context; the cost of ensuring accuracy and legal defensibility would likely exceed the cost of human expert consultation directly. |
| Cost vs. human wage | claude-sonnet-5 | 1/5 | AI cannot substitute for the human expert consultation itself, so there is no meaningful AI cost basis to compare against the human expert's wage for this specific interactive task. |
| Technical feasibility today | claude-haiku-4-5-20251001 | 1/5 | No deployed product reliably performs expert-to-expert forensic consultation and evidence interpretation at the quality required for legal/investigative contexts. AI can assist with data retrieval or summarization but cannot replace the specialized expert dialogue this task demands. |
| Technical feasibility today | claude-sonnet-5 | 1/5 | No deployed product conducts multi-expert forensic consultations or interprets evidence collaboratively across specialties; this remains firmly in human professional practice. |
Related occupations — Life, Physical & Social Science
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.