Medical and Clinical Laboratory Technicians
29-2012.00Rank #609 of 923 scored · top 66% by substitution
Perform routine medical laboratory tests for the diagnosis, treatment, and prevention of disease. May work under the supervision of a medical technologist.
Sub-scores
0–100 · band = confidence interval from rater disagreement
Substitution23
Exposure23
Augmentation57
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
13 rated tasks, binned by substitution score.
Position among all scored occupations
Distribution of 923 occupation scores; the marker is this occupation.
Tasks with substitution ≥ 70
0%
Run 1.0.0-draft.1 · computed 2026-08-05 · rater panel: claude-sonnet-5, claude-haiku-4-5-20251001 · intervals span rater disagreement.
Why this score
The five weighted dimensions of the composite, averaged across this occupation's tasks (importance-weighted, panel mean). Exact weights and formulas: /api/v1/methodology.
Task automatabilityw 35%22
panel mean rating 1.9/5 → substitution pressure 22/100
Technical feasibility todayw 20%24
panel mean rating 1.9/5 → substitution pressure 24/100
Cost vs. human wagew 15%22
panel mean rating 1.9/5 → substitution pressure 22/100
Adoption barriersw 20%inverted — strong barriers lower the score19
panel mean rating 4.2/5 (barrier strength) → substitution pressure 19/100
Sector adoption velocityw 10%31
panel mean rating 2.2/5 → substitution pressure 31/100
Task breakdown (13 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.
Conduct chemical analyses of body fluids, such as blood or urine, using microscope or automatic analyzer to detect abnormalities or diseases and enter findings into computer.
51CI 32–70 · exposure 55 · augmentation 75 · importance 4.8/5 · click for rater detail
Conduct chemical analyses of body fluids, such as blood or urine, using microscope or automatic analyzer to detect abnormalities or diseases and enter findings into computer.
51| Dimension | Model | Rating | Rationale |
|---|---|---|---|
| Sector adoption velocity | claude-haiku-4-5-20251001 | 4/5 | Large hospital systems and reference labs have rapidly adopted automated analyzers and AI result interpretation over the past decade; adoption is especially fast in well-funded clinical settings and regional labs. Smaller clinics lag, but the trajectory in the core sectors (hospitals, diagnostic companies) is steep and measurable. |
| Sector adoption velocity | claude-sonnet-5 | 3/5 | Clinical laboratories have adopted automated analyzers extensively over decades, but adoption of newer AI-driven diagnostic tools is more cautious due to regulatory approval processes and patient safety concerns. |
| Augmentation potential | claude-haiku-4-5-20251001 | 4/5 | AI analyzers substantially boost technician productivity by automating routine chemical analysis and initial interpretation, allowing them to focus on exception handling, complex cases, and quality oversight. The human remains in the loop, and their throughput and accuracy both improve. |
| Augmentation potential | claude-sonnet-5 | 4/5 | Automated analyzers and AI-assisted image analysis significantly speed up throughput and flag abnormalities for technician review, meaningfully boosting productivity while humans remain responsible for verification and judgment calls. |
| Task automatability | claude-haiku-4-5-20251001 | 4/5 | Automatic analyzers already perform the chemical analysis portion at scale with high sensitivity, and AI can reliably interpret results and flag abnormalities. Data entry into computers is fully automatable. The primary remaining human involvement is specimen preparation, quality control validation, and handling edge cases—achievable workflows could save >50% of technician time on routine cases. |
| Task automatability | claude-sonnet-5 | 2/5 | Automated analyzers already handle much of the chemical analysis, but sample handling, quality control, microscopic review of atypical specimens, and troubleshooting still require human technicians; full end-to-end automation is not yet standard for the whole workflow. |
| Adoption barriers | claude-haiku-4-5-20251001 | 3/5 | Clinical laboratories are regulated (CLIA, CAP) and technician licensure is required in most U.S. states, but these requirements constrain *personnel*, not necessarily the *automation itself*. Labs can adopt AI-assisted analysis while keeping a licensed tech on-site for oversight and exceptions—moderate friction rather than a hard legal bar against automation. |
| Adoption barriers | claude-sonnet-5 | 4/5 | Clinical lab testing is heavily regulated (CLIA, CAP accreditation) and typically requires certified personnel to perform, verify, and sign off on results, creating strong legal and liability barriers to full automation. |
| Cost vs. human wage | claude-haiku-4-5-20251001 | 4/5 | Automatic analyzers and AI interpretation cost a fraction of the fully-loaded wage of a clinical laboratory technician (~$40–50k/year), especially amortized over high-throughput labs processing hundreds of samples daily. Integration and oversight costs are modest relative to labor displacement. |
| Cost vs. human wage | claude-sonnet-5 | 2/5 | Analyzer equipment and LIS integration involve significant capital and maintenance costs, and technicians are still needed for oversight, calibration, and exception handling, so cost savings versus human labor are moderate rather than order-of-magnitude. |
| Technical feasibility today | claude-haiku-4-5-20251001 | 4/5 | Laboratory analyzers with AI-assisted interpretation are deployed in hospital and clinical settings today; systems like those using machine learning for result flagging and quality checks operate in production. Minor limitations remain in interpreting complex or ambiguous results, but the core task is demonstrably performed by mature products. |
| Technical feasibility today | claude-sonnet-5 | 3/5 | Automated clinical chemistry and hematology analyzers are mature, widely deployed products, but microscopy-based review and edge-case interpretation still rely on trained technicians rather than fully autonomous AI systems. |
Analyze and record test data to issue reports that use charts, graphs, or narratives.
41CI 37–45 · exposure 45 · augmentation 75 · importance 4.2/5 · click for rater detail
Analyze and record test data to issue reports that use charts, graphs, or narratives.
41| Dimension | Model | Rating | Rationale |
|---|---|---|---|
| Sector adoption velocity | claude-haiku-4-5-20251001 | 3/5 | Clinical laboratories are moderately digitized with widespread LIS use, but AI-driven autonomous report generation remains in pilot or early deployment phases; adoption is slowed by regulatory caution, need for validation, and institutional preference for human sign-off on clinical outputs. |
| Sector adoption velocity | claude-sonnet-5 | 2/5 | Healthcare and clinical labs are historically slower adopters of AI-driven reporting tools compared to finance or tech, constrained by regulatory approval processes and legacy systems. |
| Augmentation potential | claude-haiku-4-5-20251001 | 4/5 | AI significantly assists technicians by automating data entry, generating preliminary charts and narrative templates, flagging outliers, and suggesting interpretations; these tools measurably improve productivity and reduce manual transcription burden while the technician retains critical judgment and sign-off responsibility. |
| Augmentation potential | claude-sonnet-5 | 4/5 | AI tools can significantly assist technicians by auto-generating charts, flagging anomalies, and drafting narrative summaries, while the technician still reviews and finalizes the report. |
| Task automatability | claude-haiku-4-5-20251001 | 3/5 | AI can partially automate this task: it can extract numerical data from lab instruments, generate standard charts/graphs, and draft routine narrative reports with high consistency. However, significant clinical judgment, interpretation of abnormal results, and flagging critical findings for physician review often require human oversight, limiting end-to-end automation to roughly 50% of workflow complexity. |
| Task automatability | claude-sonnet-5 | 3/5 | AI can analyze structured test data and generate charts, graphs, and narrative summaries reasonably well, but requires integration with lab information systems and validated data pipelines, and human verification remains necessary for accuracy and compliance.atab |
| Adoption barriers | claude-haiku-4-5-20251001 | 4/5 | Regulatory frameworks (CLIA, CAP) mandate that clinical laboratory results meet documented quality standards and often require a licensed technician or pathologist to sign off on critical reports; liability asymmetry and accreditation requirements create strong legal barriers to fully autonomous AI-generated clinical reports without human validation. |
| Adoption barriers | claude-sonnet-5 | 4/5 | Clinical lab results often require certified technician or pathologist review and sign-off due to liability, accreditation (CLIA), and patient safety regulations, creating strong barriers to full automation. |
| Cost vs. human wage | claude-haiku-4-5-20251001 | 3/5 | Current AI solutions (LIS integration, automated charting) reduce manual transcription and basic report generation but require significant setup, integration, and human oversight; the all-in cost is roughly comparable to a technician's loaded wage for the portions being addressed. |
| Cost vs. human wage | claude-sonnet-5 | 3/5 | Automating report generation and charting could reduce technician time on data entry and formatting, but the cost of integration, validation, and regulatory compliance keeps overall savings moderate rather than order-of-magnitude. |
| Technical feasibility today | claude-haiku-4-5-20251001 | 3/5 | Products exist (laboratory information systems with charting modules, clinical decision support tools) that perform aspects of this task reliably, but they operate within narrow scope and typically require technician validation of data quality and critical results before release, reflecting material gaps in full autonomy. |
| Technical feasibility today | claude-sonnet-5 | 2/5 | Some LIS/EHR systems include automated reporting templates and dashboards, but fully autonomous generation of clinical report narratives from raw lab data is not yet standard practice at scale in most labs. |
Examine cells stained with dye to locate abnormalities.
39CI 32–45 · exposure 42 · augmentation 75 · importance 4.5/5 · click for rater detail
Examine cells stained with dye to locate abnormalities.
39| Dimension | Model | Rating | Rationale |
|---|---|---|---|
| Sector adoption velocity | claude-haiku-4-5-20251001 | 3/5 | Healthcare is moderately digitized and early-adopter hospitals and reference labs are piloting AI-assisted screening, but widespread production deployment of autonomous cell analysis remains limited. Most labs use these tools as assistants, not replacements. |
| Sector adoption velocity | claude-sonnet-5 | 3/5 | Clinical laboratory automation is proceeding steadily (e.g., digital pathology, automated cell counters) but full AI-driven diagnostic adoption remains uneven across facilities and is subject to regulatory pacing. |
| Augmentation potential | claude-haiku-4-5-20251001 | 4/5 | AI can meaningfully assist technicians by flagging candidate abnormalities, prioritizing slides for review, and reducing scan time, thereby raising throughput and sensitivity. The human expert remains in the loop to validate and interpret findings. |
| Augmentation potential | claude-sonnet-5 | 4/5 | AI-assisted image analysis tools significantly speed up screening and flag likely abnormalities for human review, meaningfully boosting technician throughput and consistency. |
| Task automatability | claude-haiku-4-5-20251001 | 2/5 | While AI image analysis can detect some cellular abnormalities in microscopy images, this task requires nuanced judgment about morphology, context, and clinical significance that current systems handle inconsistently. Most real-world deployment still requires human expert review, preventing the ≥50% time-savings bar. |
| Task automatability | claude-sonnet-5 | 3/5 | Automated image analysis (e.g., digital cytology/hematology systems) can flag abnormalities in stained cell samples, but full end-to-end replacement including sample quality judgment and edge-case interpretation is not yet standard.' |
| Adoption barriers | claude-haiku-4-5-20251001 | 4/5 | Clinical laboratory work is heavily regulated (CLIA, CAP accreditation); diagnostic results must be validated and signed off by licensed personnel (medical technologists or pathologists). Liability for misdiagnosis and the requirement for professional accountability create strong legal barriers to full automation. |
| Adoption barriers | claude-sonnet-5 | 4/5 | Clinical lab results typically require certified/licensed personnel to review and sign off per CLIA and regulatory standards, creating strong liability and compliance barriers to full automation. |
| Cost vs. human wage | claude-haiku-4-5-20251001 | 2/5 | The cost of inference plus integration into lab workflows and mandatory human oversight is comparable to or only modestly cheaper than technician labor for this task. Integration, model maintenance, and quality assurance add overhead. |
| Cost vs. human wage | claude-sonnet-5 | 3/5 | Automated slide scanners and analyzers have high upfront capital and maintenance costs, offsetting labor savings; cost advantage is moderate rather than dramatic for many labs. |
| Technical feasibility today | claude-haiku-4-5-20251001 | 3/5 | AI-assisted microscopy analysis tools exist in production (e.g., deep learning models for detecting specific pathologies in cytology), but they typically operate as screening aids with material false-negative and false-positive rates rather than autonomous end-to-end systems. Human oversight remains standard practice. |
| Technical feasibility today | claude-sonnet-5 | 3/5 | FDA-cleared digital pathology and automated hematology analyzers (e.g., for Pap smears, blood differentials) are deployed in labs, but many still require technician confirmation and cannot handle all specimen types or abnormalities reliably. |
Analyze the results of tests or experiments to ensure conformity to specifications, using special mechanical or electrical devices.
35CI 25–45 · exposure 38 · augmentation 63 · importance 4.8/5 · click for rater detail
Analyze the results of tests or experiments to ensure conformity to specifications, using special mechanical or electrical devices.
35| Dimension | Model | Rating | Rationale |
|---|---|---|---|
| Sector adoption velocity | claude-haiku-4-5-20251001 | 3/5 | Healthcare laboratories have invested in LIS and digital workflows, but adoption of AI-powered autonomous analysis is in pilot and early production phases rather than widespread deployment; progress is faster in large hospital systems than smaller labs. |
| Sector adoption velocity | claude-sonnet-5 | 2/5 | Healthcare and clinical lab settings adopt automation slowly due to regulatory validation requirements and safety-critical nature of diagnostic testing. |
| Augmentation potential | claude-haiku-4-5-20251001 | 4/5 | AI significantly assists technicians by automatically flagging abnormal results, detecting quality failures, and highlighting patterns requiring review, thereby reducing manual scanning and freeing technicians to focus on exceptions and judgment-heavy validation. |
| Augmentation potential | claude-sonnet-5 | 3/5 | AI-assisted analyzers and software can flag anomalies or out-of-range results, helping technicians work faster, though human confirmation remains essential. |
| Task automatability | claude-haiku-4-5-20251001 | 3/5 | AI can automate parts of result analysis—flagging outliers, comparing against reference ranges, and detecting obvious deviations from specifications—but clinical validation and judgment about specimen quality issues typically require human oversight, achieving roughly 50% time savings with setup. |
| Task automatability | claude-sonnet-5 | 2/5 | Interpreting instrument readouts against specifications involves some pattern recognition AI can assist with, but the physical operation of mechanical/electrical devices and final quality judgment on samples remains largely human-performed today. |
| Adoption barriers | claude-haiku-4-5-20251001 | 4/5 | Clinical laboratory results have regulatory oversight (CLIA, CAP) and liability requirements; a licensed technologist or pathologist must typically verify and sign off on results, creating a strong legal and professional barrier to full automation. |
| Adoption barriers | claude-sonnet-5 | 4/5 | Clinical lab results often require certified technician review and sign-off under regulatory frameworks (e.g., CLIA), creating strong legal and liability barriers to full automation. |
| Cost vs. human wage | claude-haiku-4-5-20251001 | 3/5 | AI-enabled analysis software and LIS integration cost is roughly comparable to the fractionated time technicians spend on repetitive result review once development and integration overhead is accounted for. |
| Cost vs. human wage | claude-sonnet-5 | 2/5 | Lab equipment with embedded analytics already carries automation costs, but human technicians are still required for calibration, troubleshooting and verification, keeping AI-only substitution costlier than assumed. |
| Technical feasibility today | claude-haiku-4-5-20251001 | 3/5 | Laboratory information systems (LIS) with rule-based quality checks and some AI-enabled anomaly detection exist in production, but they require significant human review; no fully autonomous end-to-end result validation system is reliably deployed at scale without technician sign-off. |
| Technical feasibility today | claude-sonnet-5 | 2/5 | Some lab analyzers include automated flagging and QC software, but full autonomous analysis and conformity determination without technician oversight is not standard deployed practice. |
Test raw materials, processes, or finished products to determine quality or quantity of materials or characteristics of a substance.
26CI 25–28 · exposure 25 · augmentation 63 · importance 4.4/5 · click for rater detail
Test raw materials, processes, or finished products to determine quality or quantity of materials or characteristics of a substance.
26| Dimension | Model | Rating | Rationale |
|---|---|---|---|
| Sector adoption velocity | claude-haiku-4-5-20251001 | 2/5 | While hospitals and larger labs are piloting AI-assisted analysis tools, adoption remains limited; most clinical and industrial laboratories still rely on human technicians for routine testing, with AI used only as a secondary tool in specialized applications. |
| Sector adoption velocity | claude-sonnet-5 | 3/5 | Clinical labs have adopted automated analyzers and LIS/AI-assisted flagging systems for years, but full AI-driven testing pipelines remain limited to narrow high-volume tests rather than broad deployment. |
| Augmentation potential | claude-haiku-4-5-20251001 | 4/5 | AI excels at assisting technicians through automated image analysis, pattern recognition in test results, and quality control flagging, which can speed interpretation and reduce manual review time while the technician retains responsibility for final validation and sign-off. |
| Augmentation potential | claude-sonnet-5 | 3/5 | AI-assisted instruments, quality control flagging, and pattern recognition (e.g., in microscopy or chemistry panels) help technicians work faster and catch anomalies, though core physical testing remains human-performed. |
| Task automatability | claude-haiku-4-5-20251001 | 2/5 | While AI can analyze some laboratory data (e.g., image-based microscopy or spectroscopy results), the full task involves physical specimen handling, instrument operation, and interpretation of complex material properties that remain heavily dependent on human execution and judgment in most laboratory settings. |
| Task automatability | claude-sonnet-5 | 2/5 | Physical specimen handling, running assays, and operating lab instruments require manual manipulation and equipment operation that current AI cannot perform end-to-end without robotics; AI can assist with data interpretation but not the full physical testing workflow. |
| Adoption barriers | claude-haiku-4-5-20251001 | 4/5 | Laboratory testing is heavily regulated (CLIA, CAP, ISO standards) and often requires licensed or certified technicians to perform and certify results; liability for incorrect test results creates significant legal and compliance barriers to full automation. |
| Adoption barriers | claude-sonnet-5 | 4/5 | Clinical lab testing is heavily regulated (CLIA, CAP, FDA) requiring certified personnel and validated methods, with significant liability for diagnostic errors, creating strong barriers to full automation. |
| Cost vs. human wage | claude-haiku-4-5-20251001 | 2/5 | AI infrastructure for laboratory analysis (imaging systems, software, integration) is capital-intensive, and the ongoing need for human technician oversight, instrument maintenance, and sample preparation means total cost per test remains comparable to or exceeds direct labor. |
| Cost vs. human wage | claude-sonnet-5 | 2/5 | Specialized lab automation equipment has high upfront capital and maintenance costs, and still requires technician oversight, so total cost is not dramatically below skilled technician wages for many tests. |
| Technical feasibility today | claude-haiku-4-5-20251001 | 2/5 | Some narrow AI applications exist for image analysis in pathology or quality defect detection, but deployed systems typically require human technicians to prepare samples, operate instruments, and validate results; end-to-end automation with reliable performance across diverse material types is not standard in production labs. |
| Technical feasibility today | claude-sonnet-5 | 2/5 | Automated analyzers and lab information systems exist and are widely deployed for specific assays, but general-purpose AI performing the full range of qualitative/quantitative testing across substances is not a mature deployed product. |
Conduct blood tests for transfusion purposes and perform blood counts.
24CI 20–28 · exposure 25 · augmentation 63 · importance 4.9/5 · click for rater detail
Conduct blood tests for transfusion purposes and perform blood counts.
24| Dimension | Model | Rating | Rationale |
|---|---|---|---|
| Sector adoption velocity | claude-haiku-4-5-20251001 | 3/5 | Clinical laboratories have adopted automated analyzers widely, but adoption is incremental and bottlenecked by regulation, capital cost, and the requirement for human oversight. Most labs are in a hybrid state—machines handle routine counts, but technicians remain central to the workflow. |
| Sector adoption velocity | claude-sonnet-5 | 2/5 | Clinical laboratories have adopted automated analyzers over decades, but adoption of newer AI/agentic systems for this specific regulated task is slow due to healthcare's cautious regulatory environment. |
| Augmentation potential | claude-haiku-4-5-20251001 | 4/5 | AI-assisted blood analysis platforms significantly enhance technician productivity by automating counts, flagging abnormalities, and reducing manual microscopy time. These systems allow technicians to focus on verification and complex cases, substantially raising output per technician. |
| Augmentation potential | claude-sonnet-5 | 3/5 | Automated analyzers and lab information systems meaningfully speed up blood counts and flag abnormal results, augmenting technician throughput and accuracy while humans retain oversight and final verification. |
| Task automatability | claude-haiku-4-5-20251001 | 2/5 | While AI can assist with analyzing blood count results and flagging transfusion compatibility data, the actual blood draw, specimen handling, and equipment operation remain manual. Current systems cannot perform the full end-to-end task (collection through interpretation and quality assurance) with ≥50% time savings at equal quality. |
| Task automatability | claude-sonnet-5 | 2/5 | Blood counts are already largely automated by lab analyzers, but this task also includes physical sample handling, transfusion compatibility testing, and quality verification requiring manual dexterity and judgment that current AI cannot end-to-end perform.}, |
| Adoption barriers | claude-haiku-4-5-20251001 | 4/5 | Blood transfusion and testing are highly regulated (FDA, CLIA, CAP standards), and most jurisdictions require a licensed clinical laboratory technician to perform, certify, or sign off on transfusion-critical results. Patient safety liability and regulatory mandates create substantial legal barriers to full automation. |
| Adoption barriers | claude-sonnet-5 | 5/5 | Transfusion testing is safety-critical and heavily regulated, requiring certified/licensed laboratory personnel to perform and verify results due to severe risk of harm from errors. |
| Cost vs. human wage | claude-haiku-4-5-20251001 | 2/5 | While automated hematology analyzers reduce per-sample labor, initial capital cost, maintenance, reagents, and required technician oversight mean the all-in cost is comparable to or exceeds a technician's loaded wage for the same throughput and accuracy standards. |
| Cost vs. human wage | claude-sonnet-5 | 2/5 | Lab automation equipment is expensive capital investment with ongoing calibration and human oversight needs, so while throughput improves, the all-in cost relative to a technician is not dramatically cheaper given equipment and maintenance costs. |
| Technical feasibility today | claude-haiku-4-5-20251001 | 2/5 | Blood analyzers exist that automate counting and basic flagging, but they require trained technician operation, specimen preparation, and quality control oversight. No deployed system performs the complete transfusion testing and blood counting task autonomously; human judgment and intervention remain essential. |
| Technical feasibility today | claude-sonnet-5 | 2/5 | Automated hematology analyzers and blood typing instruments are deployed widely, but these are specialized lab hardware/software rather than general AI, and full task including specimen handling and cross-matching still requires human technicians. |
Prepare standard volumetric solutions or reagents to be combined with samples, following standardized formulas or experimental procedures.
23CI 21–25 · exposure 25 · augmentation 50 · importance 4.6/5 · click for rater detail
Prepare standard volumetric solutions or reagents to be combined with samples, following standardized formulas or experimental procedures.
23| Dimension | Model | Rating | Rationale |
|---|---|---|---|
| Sector adoption velocity | claude-haiku-4-5-20251001 | 2/5 | Clinical laboratories are traditionally conservative and highly regulated; adoption of robotic reagent prep is limited to large academic medical centers and reference labs with high-throughput needs, not mainstream practice. Most labs continue manual preparation by trained technicians. |
| Sector adoption velocity | claude-sonnet-5 | 2/5 | Clinical laboratories adopt automation slowly due to regulatory validation requirements, though large reference labs have automated some solution preparation; most facilities still rely on manual prep. |
| Augmentation potential | claude-haiku-4-5-20251001 | 3/5 | AI could usefully assist by automatically calculating precise volumes from formulas, checking procedural compliance, flagging concentration ranges, and generating tracking documentation, raising technician efficiency on the calculation and documentation phases while the human remains responsible for actual preparation and QC. |
| Augmentation potential | claude-sonnet-5 | 3/5 | AI can assist with calculating dilutions, tracking inventory, generating standardized protocols, and flagging errors, but the physical preparation itself remains human-performed. |
| Task automatability | claude-haiku-4-5-20251001 | 2/5 | While AI could theoretically assist with calculating volumes and flagging procedural steps, the physical manipulation of solutions, precise pipetting, temperature control, and real-time quality verification require laboratory equipment and human dexterity that current AI systems cannot perform end-to-end. Automation exists for specific high-throughput chemistry, but not as a general off-the-shelf solution for standard reagent preparation. |
| Task automatability | claude-sonnet-5 | 2/5 | Preparing precise reagents and volumetric solutions requires physical manipulation, pipetting, weighing, and dilution that current AI cannot perform end-to-end; only documentation/calculation portions are automatable.' |
| Adoption barriers | claude-haiku-4-5-20251001 | 4/5 | Clinical laboratory reagent preparation is subject to CLIA/CAP regulations and ISO 15189 standards that require validated procedures and documented human oversight or certification of the preparation process. Liability for incorrect reagent concentration directly affects patient results, creating strong regulatory and safety barriers to full automation without certified human sign-off. |
| Adoption barriers | claude-sonnet-5 | 4/5 | Clinical lab reagent preparation is subject to strict regulatory quality control (CLIA, CAP) requiring documented procedures and often specific personnel qualifications, creating substantial compliance barriers to full automation. |
| Cost vs. human wage | claude-haiku-4-5-20251001 | 1/5 | Specialized robotic liquid handling systems capable of reagent preparation are capital-intensive (hundreds of thousands of dollars) and have high maintenance costs, making them more expensive than a technician's loaded wage for most labs, especially smaller facilities. |
| Cost vs. human wage | claude-sonnet-5 | 2/5 | Automated liquid handling equipment is expensive to purchase, validate, and maintain, often costing more than technician labor for lower-volume or specialized reagent prep tasks. |
| Technical feasibility today | claude-haiku-4-5-20251001 | 2/5 | No deployed AI product reliably performs full reagent preparation autonomously in production clinical labs today. Robotic liquid handlers exist but require custom integration per protocol and are expensive specialized equipment, not generalizable AI systems. Manual preparation by trained technicians remains standard. |
| Technical feasibility today | claude-sonnet-5 | 2/5 | Lab automation systems (liquid handlers, automated reagent preparation stations) exist but require significant capital investment and are not universally deployed for this exact task in most clinical labs; AI itself doesn't perform physical prep. |
Set up, maintain, calibrate, clean, and test sterility of medical laboratory equipment.
16CI 7–25 · exposure 13 · augmentation 38 · importance 4.7/5 · click for rater detail
Set up, maintain, calibrate, clean, and test sterility of medical laboratory equipment.
16| Dimension | Model | Rating | Rationale |
|---|---|---|---|
| Sector adoption velocity | claude-haiku-4-5-20251001 | 2/5 | Healthcare and laboratory settings adopt AI and automation slowly due to regulatory constraints, risk aversion, and the need for validated, auditable processes. Equipment maintenance remains largely manual and technician-driven, with adoption of AI-assisted tools limited to scheduling and documentation rather than the core technical work. |
| Sector adoption velocity | claude-sonnet-5 | 2/5 | Clinical laboratory operations are moderately digitized but physical equipment maintenance and sterility testing remain manual, with AI adoption in this specific physical task lagging behind digital/diagnostic AI uses in healthcare. |
| Augmentation potential | claude-haiku-4-5-20251001 | 3/5 | AI can usefully assist by automating maintenance schedules, flagging calibration drift patterns, generating compliance documentation, and reminding technicians of sterility protocols. However, the human technician must remain the decision-maker and executor for the physical and verification tasks themselves. |
| Augmentation potential | claude-sonnet-5 | 2/5 | AI can offer some assistance via automated calibration logs, predictive maintenance alerts, or digital tracking systems, but it does not meaningfully transform the physical execution of setup, cleaning, or sterility testing. |
| Task automatability | claude-haiku-4-5-20251001 | 2/5 | While AI could assist with maintenance scheduling and some calibration documentation, the physical setup, cleaning, and hands-on sterility testing of laboratory equipment require manual dexterity and in-situ environmental assessment that current AI systems cannot perform end-to-end. Robotics exist for some repetitive lab tasks, but general equipment maintenance remains largely manual. |
| Task automatability | claude-sonnet-5 | 1/5 | This is a hands-on physical task requiring manipulation of lab equipment, sterility testing, and calibration procedures that current AI cannot perform without robotic embodiment, which is not generally available. |
| Adoption barriers | claude-haiku-4-5-20251001 | 4/5 | Regulatory frameworks (FDA, CLIA, ISO standards) often require documented human accountability for equipment calibration and sterility assurance. Many laboratory quality protocols mandate that a licensed or certified technician verify calibration and sterility, creating a hard legal and compliance barrier to full automation. |
| Adoption barriers | claude-sonnet-5 | 4/5 | Sterility testing and equipment calibration in clinical labs are subject to strict regulatory and accreditation standards (e.g., CLIA, CAP) requiring qualified personnel to perform and document these procedures, creating strong barriers to automation. |
| Cost vs. human wage | claude-haiku-4-5-20251001 | 2/5 | Current automation solutions (where they exist) are costly to integrate and maintain, and the task requires frequent human oversight and intervention. The loaded cost of specialized robotics or AI systems for laboratory equipment maintenance exceeds the wage of skilled technicians, especially given low task volume per facility. |
| Cost vs. human wage | claude-sonnet-5 | 1/5 | There is no AI substitute performing this physical task, so AI cost is not comparable—human labor remains the only viable option, making AI effectively infinitely costlier for this specific function. |
| Technical feasibility today | claude-haiku-4-5-20251001 | 2/5 | No deployed AI product reliably handles the full spectrum of equipment setup, calibration, and sterility verification independently. Some lab management software assists with scheduling and record-keeping, but the core technical work—calibrating instruments, physically cleaning equipment, performing sterility tests—remains human-performed in production settings. |
| Technical feasibility today | claude-sonnet-5 | 1/5 | No deployed AI product performs physical equipment setup, calibration, cleaning, or sterility testing in clinical labs today; this remains entirely manual work. |
Obtain specimens, cultivating, isolating, and identifying microorganisms for analysis.
16CI 7–25 · exposure 13 · augmentation 50 · importance 4.7/5 · click for rater detail
Obtain specimens, cultivating, isolating, and identifying microorganisms for analysis.
16| Dimension | Model | Rating | Rationale |
|---|---|---|---|
| Sector adoption velocity | claude-haiku-4-5-20251001 | 2/5 | Laboratory settings show slow adoption of AI-driven automation for specimen work; most labs still rely on manual microscopy and culturing. Pilots exist in large academic medical centers, but production deployment of fully autonomous identification or handling remains limited, particularly in smaller/rural labs. |
| Sector adoption velocity | claude-sonnet-5 | 2/5 | Clinical laboratory settings adopt automation (e.g., automated culture analyzers) but AI-driven full task replacement is slow given physical and regulatory constraints; robotics-assisted sample handling exists but is not AI-driven broadly yet. |
| Augmentation potential | claude-haiku-4-5-20251001 | 3/5 | AI-powered microscopy and image recognition systems meaningfully assist technicians in organism identification and classification, reducing review time and improving consistency. However, specimen acquisition and culturing still require human decisions and hands-on work, limiting overall productivity gain. |
| Augmentation potential | claude-sonnet-5 | 3/5 | AI can assist with image-based microorganism identification and interpretation of culture results, aiding technicians in classification and diagnostic decision support, though physical specimen work remains manual. |
| Task automatability | claude-haiku-4-5-20251001 | 2/5 | While AI can assist with image analysis and identification of microorganisms from cultured samples, obtaining and cultivating specimens requires physical manipulation, sterile technique, and contextual decision-making that current automation cannot reliably perform end-to-end. AI imaging systems exist but cannot independently perform sample preparation, culturing conditions, or decision trees for what to culture next. |
| Task automatability | claude-sonnet-5 | 1/5 | This is a hands-on physical and biological task requiring specimen collection, sterile technique, and physical manipulation of cultures—current AI systems cannot perform physical lab work. |
| Adoption barriers | claude-haiku-4-5-20251001 | 4/5 | Laboratory testing is heavily regulated (CLIA, CAP, ISO standards) and typically requires licensed clinical laboratory personnel to perform and validate specimen handling and organism identification. Results often require professional sign-off and chain-of-custody documentation, creating hard barriers to full automation. |
| Adoption barriers | claude-sonnet-5 | 4/5 | Clinical lab work is regulated (CLIA certification, licensure in many jurisdictions) and requires trained personnel for specimen handling and biosafety, creating strong regulatory and safety barriers to automation. |
| Cost vs. human wage | claude-haiku-4-5-20251001 | 2/5 | Automated specimen handling and culture equipment exists but remains expensive; AI image analysis for identification has modest inference cost but integration with lab automation is non-trivial. The all-in cost (hardware, software, integration, validation, oversight) typically exceeds the cost of a trained technician for most labs. |
| Cost vs. human wage | claude-sonnet-5 | 1/5 | AI cannot substitute for the physical labor and equipment interaction involved, so there is no meaningful AI cost comparison—human labor remains the only option. |
| Technical feasibility today | claude-haiku-4-5-20251001 | 2/5 | AI-assisted microscopy and image recognition for organism identification exist in research and some lab settings, but no deployed product reliably performs the full pipeline of specimen acquisition, cultivation, isolation, and identification without substantial human oversight and manual handling. Current systems require expert human judgment on multiple steps. |
| Technical feasibility today | claude-sonnet-5 | 1/5 | No deployed AI product can obtain specimens or physically cultivate/isolate microorganisms; this remains entirely a manual laboratory procedure performed by trained technicians. |
Perform medical research to further control or cure disease.
11CI 3–20 · exposure 8 · augmentation 63 · importance 4.4/5 · click for rater detail
Perform medical research to further control or cure disease.
11| Dimension | Model | Rating | Rationale |
|---|---|---|---|
| Sector adoption velocity | claude-haiku-4-5-20251001 | 2/5 | While AI adoption in research support is growing (literature mining, data analysis), actual displacement of research personnel performing original disease research is minimal. Most organizations still rely on human researchers at the core of discovery, with AI serving as an assistant tool rather than a replacement. |
| Sector adoption velocity | claude-sonnet-5 | 2/5 | Healthcare and clinical lab settings have historically been slower to adopt AI for hands-on research tasks compared to purely digital/information sectors, though AI-assisted analysis is growing. |
| Augmentation potential | claude-haiku-4-5-20251001 | 3/5 | AI can meaningfully assist medical researchers with literature synthesis, statistical analysis, data visualization, and hypothesis generation from large datasets, improving researcher productivity on parts of the research workflow. However, the augmentation is limited to support functions rather than transforming the core task of designing and executing novel research. |
| Augmentation potential | claude-sonnet-5 | 4/5 | AI can significantly help technicians with literature reviews, data analysis, pattern recognition in experimental results, and drafting research documentation, improving overall research productivity. |
| Task automatability | claude-haiku-4-5-20251001 | 1/5 | Medical research aimed at disease control or cure requires hypothesis formulation, experimental design, interpretation of novel biological data, and creative scientific reasoning. Current AI systems cannot autonomously conduct original research that meets peer-review standards or produces publishable scientific advances. |
| Task automatability | claude-sonnet-5 | 2/5 | Medical research involves hands-on experimentation, hypothesis generation, and lab work that AI cannot fully replace, though literature review and data analysis portions can be assisted.' |
| Adoption barriers | claude-haiku-4-5-20251001 | 5/5 | Medical research is heavily regulated by IRBs, funding agencies, and publication standards. Human researchers must design studies, obtain ethical approval, and take intellectual and legal responsibility for findings. Regulatory and liability barriers, combined with the requirement for human scientific judgment and accountability, create strong legal and institutional barriers to full automation. |
| Adoption barriers | claude-sonnet-5 | 4/5 | Medical research involving human subjects or clinical applications is heavily regulated, requiring institutional review, credentialed oversight, and accountability that AI cannot assume. |
| Cost vs. human wage | claude-haiku-4-5-20251001 | 1/5 | Medical researchers command high salaries, and the value of original disease research is substantial. AI tools that support research are far cheaper than researchers themselves, but the task as stated—performing research to further control or cure disease—requires expertise and institutional oversight that remains more expensive to deploy via AI than to employ skilled humans. |
| Cost vs. human wage | claude-sonnet-5 | 2/5 | AI can cheaply assist with data analysis or literature review, but the physical experimentation and specialized lab skills still require human technicians, limiting overall cost savings. |
| Technical feasibility today | claude-haiku-4-5-20251001 | 1/5 | No deployed product performs original medical research autonomously. While AI can assist with literature review and data analysis, the core task of designing novel experiments and conducting original research to advance disease treatment requires human researchers and is not performed end-to-end by AI systems in production. |
| Technical feasibility today | claude-sonnet-5 | 1/5 | No deployed product autonomously conducts medical research; AI tools exist for literature search and data analysis but not for the full research task performed by lab technicians. |
Consult with a pathologist to determine a final diagnosis when abnormal cells are found.
7CI 3–11 · exposure 5 · augmentation 63 · importance 4.5/5 · click for rater detail
Consult with a pathologist to determine a final diagnosis when abnormal cells are found.
7| Dimension | Model | Rating | Rationale |
|---|---|---|---|
| Sector adoption velocity | claude-haiku-4-5-20251001 | 2/5 | While pathology labs are adopting AI for image analysis and screening triage, the final diagnosis consultation step remains a human-expert bottleneck; adoption of AI in this exact task is slow because the regulatory and liability structure preserves the pathologist's gatekeeping role. |
| Sector adoption velocity | claude-sonnet-5 | 2/5 | Healthcare diagnostics adoption of AI is growing but remains slow and heavily regulated, especially for final diagnostic decision-making rather than image triage or screening support. |
| Augmentation potential | claude-haiku-4-5-20251001 | 4/5 | AI can substantially assist by pre-analyzing slides, highlighting suspicious areas, retrieving similar cases, and summarizing relevant literature before consultation, meaningfully improving the pathologist's speed and confidence in reaching a final diagnosis while keeping the expert in the loop. |
| Augmentation potential | claude-sonnet-5 | 3/5 | AI tools (e.g., digital pathology image analysis, decision support) can assist by flagging abnormal cells or suggesting differentials, aiding the consultation but not replacing the clinical dialogue. |
| Task automatability | claude-haiku-4-5-20251001 | 1/5 | This task requires expert pathologist consultation to render a final diagnosis on abnormal cells—a human judgment call grounded in medical expertise, pattern recognition across complex histological data, and clinical context. Current AI cannot autonomously conduct this consultation or override/replace the pathologist's role in diagnosis validation. |
| Task automatability | claude-sonnet-5 | 1/5 | This requires a licensed pathologist's clinical judgment and a consultative dialogue about diagnosis; AI cannot perform this end-to-end interaction and decision today.5 |
| Adoption barriers | claude-haiku-4-5-20251001 | 5/5 | Pathology diagnosis is legally and clinically mandated to be performed or signed off by a licensed pathologist; liability, regulatory oversight (CAP, CLIA), and standard of care all require human expert judgment, creating hard barriers to automation. |
| Adoption barriers | claude-sonnet-5 | 5/5 | Final diagnosis in pathology is a licensed medical act with strict regulatory, liability, and credentialing requirements, making this one of the most protected clinical tasks. |
| Cost vs. human wage | claude-haiku-4-5-20251001 | 2/5 | AI-assisted screening can reduce initial review workload, but the consultation itself requires a licensed pathologist's time, and integration costs, oversight, and liability management make the all-in cost similar to or higher than the human expert's time for this specific consultation step. |
| Cost vs. human wage | claude-sonnet-5 | 1/5 | There is no AI system replacing this consultation task, so cost comparison favors the human process entirely; any AI tool would be additive cost, not a substitute. |
| Technical feasibility today | claude-haiku-4-5-20251001 | 2/5 | AI systems exist to flag abnormalities and assist with image analysis, but no deployed product independently performs end-to-end diagnosis consultation with a pathologist; the system must integrate into existing workflows and the pathologist retains decision authority in clinical practice. |
| Technical feasibility today | claude-sonnet-5 | 1/5 | No deployed product substitutes for the pathologist consultation itself; AI diagnostic aids exist but the actual consult-and-decide step remains a human-to-human clinical process. |
Supervise or instruct other technicians or laboratory assistants.
4CI 0–7 · exposure 0 · augmentation 38 · importance 4.1/5 · click for rater detail
Supervise or instruct other technicians or laboratory assistants.
4| Dimension | Model | Rating | Rationale |
|---|---|---|---|
| Sector adoption velocity | claude-haiku-4-5-20251001 | 1/5 | Supervisory roles have not seen adoption of autonomous AI in laboratory settings; adoption remains near-zero because the role is interpersonal, mission-critical, and legally tied to a specific individual's licensure and liability. |
| Sector adoption velocity | claude-sonnet-5 | 2/5 | Clinical laboratories are moderately digitized but management/supervisory functions see little AI adoption; most AI use is in diagnostics, not personnel supervision. |
| Augmentation potential | claude-haiku-4-5-20251001 | 2/5 | AI could assist with scheduling, documentation, or performance data summaries, but these are peripheral to the core supervisory task of directing, mentoring, and evaluating personnel. Assistance is limited and not central to the role. |
| Augmentation potential | claude-sonnet-5 | 3/5 | AI tools can help schedule training, track performance metrics, or generate instructional materials, providing moderate assistance to a supervisor without replacing the interpersonal role. |
| Task automatability | claude-haiku-4-5-20251001 | 1/5 | Supervising and instructing lab personnel requires real-time judgment, conflict resolution, performance assessment, and nuanced communication that current AI systems cannot perform end-to-end. This task inherently involves human-to-human relationships and adaptive feedback that exceed what automation can deliver today. |
| Task automatability | claude-sonnet-5 | 1/5 | Supervision and instruction of staff require interpersonal leadership, real-time feedback, and accountability that AI cannot perform end-to-end today.' |
| Adoption barriers | claude-haiku-4-5-20251001 | 5/5 | Laboratory supervision is typically a formally designated role with legal accountability for quality control, safety compliance, and staff conduct. Organizations require a licensed or certified human supervisor to bear responsibility, creating a hard barrier to automation. |
| Adoption barriers | claude-sonnet-5 | 4/5 | Supervisory roles typically require credentialed, experienced personnel accountable for lab quality and compliance, creating strong organizational and regulatory barriers to automation. |
| Cost vs. human wage | claude-haiku-4-5-20251001 | 1/5 | Replacing a supervisor with AI would require continuous human oversight to validate decisions, making the all-in cost of such a system higher than the loaded wage of an actual supervisor who can make autonomous judgments. |
| Cost vs. human wage | claude-sonnet-5 | 1/5 | There is no AI substitute performing this task, so cost comparison favors the human entirely. |
| Technical feasibility today | claude-haiku-4-5-20251001 | 1/5 | No deployed AI product reliably performs supervisory and instructional duties for laboratory staff in production settings. While AI can draft performance notes or schedules, actual supervision—correcting mistakes, motivating staff, adapting to individual learning needs—remains outside deployed AI capabilities. |
| Technical feasibility today | claude-sonnet-5 | 1/5 | No deployed product manages or supervises human lab staff; this remains a human management function with no automation in production. |
Collect blood or tissue samples from patients, observing principles of asepsis to obtain blood sample.
0CI 0–0 · exposure 0 · augmentation 25 · importance 4.4/5 · click for rater detail
Collect blood or tissue samples from patients, observing principles of asepsis to obtain blood sample.
0| Dimension | Model | Rating | Rationale |
|---|---|---|---|
| Sector adoption velocity | claude-haiku-4-5-20251001 | 1/5 | Healthcare remains a low-automation sector for direct patient-contact procedures; adoption of autonomous phlebotomy is negligible in production, with healthcare organizations relying on trained human technicians. |
| Sector adoption velocity | claude-sonnet-5 | 1/5 | Healthcare physical-task settings show very slow adoption of automation for direct patient contact procedures compared to information-processing tasks. |
| Augmentation potential | claude-haiku-4-5-20251001 | 2/5 | AI offers minimal assistance for the core collection task itself, though digital tools for sample tracking and documentation provide marginal support to the technician's workflow rather than transforming the hands-on collection process. |
| Augmentation potential | claude-sonnet-5 | 2/5 | AI can assist with vein-finding imaging tools or sample tracking/logging, but offers minimal augmentation to the core manual collection act itself. |
| Task automatability | claude-haiku-4-5-20251001 | 1/5 | Blood and tissue collection requires direct physical contact with patients, needle insertion, and real-time adaptation to anatomy and patient response. Current AI systems have no capability to perform these hands-on procedures end-to-end. |
| Task automatability | claude-sonnet-5 | 1/5 | Physical blood/tissue collection requires fine motor manipulation, palpation, and physical contact with patients that current AI systems cannot perform end-to-end; robotic phlebotomy devices exist only in narrow experimental trials.' |
| Adoption barriers | claude-haiku-4-5-20251001 | 5/5 | Clinical sample collection is heavily regulated and requires licensed medical personnel; liability for patient harm, infection control standards, and regulatory requirements (CLIA, state licensing) create hard legal barriers to automation or substitution. |
| Adoption barriers | claude-sonnet-5 | 5/5 | Invasive sample collection is a regulated clinical procedure typically requiring licensed personnel, with significant liability and patient safety concerns preventing non-human substitution. |
| Cost vs. human wage | claude-haiku-4-5-20251001 | 1/5 | AI robotics for phlebotomy exist only in research or extremely limited pilot deployments; the cost of specialized hardware, calibration, and oversight far exceeds the labor cost of a trained technician performing the task. |
| Cost vs. human wage | claude-sonnet-5 | 1/5 | Any robotic sampling hardware would require expensive specialized equipment, maintenance, and human oversight, making it costlier than a trained technician for this task today. |
| Technical feasibility today | claude-haiku-4-5-20251001 | 1/5 | No deployed AI product can autonomously collect blood or tissue samples from patients. This remains a purely human task requiring trained laboratory technicians. |
| Technical feasibility today | claude-sonnet-5 | 1/5 | No deployed product reliably performs venipuncture or tissue sampling in general clinical practice today; automated blood draw devices remain research/pilot stage with limited scope. |
Related occupations — Healthcare Practitioners & Technical
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.