Molecular and Cellular Biologists

19-1029.02
Median wage $98,920/yr55,850 employed (US)Rank #422 of 923 scored · top 46% by substitution

Research and study cellular molecules and organelles to understand cell function and organization.

Sub-scores

0–100 · band = confidence interval from rater disagreement

Substitution29
Exposure25
Augmentation65

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

22 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%26

panel mean rating 2.1/5 → substitution pressure 26/100

Technical feasibility todayw 20%23

panel mean rating 1.9/5 → substitution pressure 23/100

Cost vs. human wagew 15%28

panel mean rating 2.1/5 → substitution pressure 28/100

Adoption barriersw 20%inverted — strong barriers lower the score40

panel mean rating 3.4/5 (barrier strength) → substitution pressure 40/100

Sector adoption velocityw 10%31

panel mean rating 2.2/5 → substitution pressure 31/100

Task breakdown (22 tasks)

Substitution pressure per task, weighted by O*NET importance in the composite. Expand a task for the full rater audit trail — every rating, every model, every rationale.

Maintain accurate laboratory records and data.

61

CI 48–74 · exposure 62 · augmentation 88 · importance 4.7/5 · click for rater detail

Panel ratings per dimension for this task
DimensionModelRatingRationale
Sector adoption velocityclaude-haiku-4-5-202510014/5Large pharmaceutical, biotech, and academic institutions with high-throughput labs have rapidly adopted LIMS with AI-assisted data capture and automated workflows; early-stage labs and small institutions lag, but the trend in information-intensive biomedical research is accelerating adoption.
Sector adoption velocityclaude-sonnet-52/5Academic and biotech labs are gradually adopting digital ELN and AI tools, but adoption is slower than in software-centric industries due to funding constraints, legacy paper practices, and low digitization in many labs.
Augmentation potentialclaude-haiku-4-5-202510015/5AI augments biologists by auto-populating records, flagging missing or anomalous data, auto-generating compliant summaries, and cross-linking related experiments, which substantially raises their speed and accuracy while keeping them in oversight and validation roles.
Augmentation potentialclaude-sonnet-54/5AI tools can meaningfully assist by auto-populating templates, flagging inconsistencies, summarizing data, and improving searchability, significantly aiding scientists while they remain responsible for accuracy.
Task automatabilityclaude-haiku-4-5-202510014/5Current AI systems can reliably capture, organize, and structure laboratory records from notes, images, and instrument outputs with minimal human intervention, achieving >50% time savings through automated data entry, timestamp logging, and metadata tagging. Some domain-specific validation (e.g., unit conversion, equipment calibration flags) requires oversight, but the bulk of record maintenance is automatable end-to-end.
Task automatabilityclaude-sonnet-53/5AI can assist with structuring, transcribing, and organizing lab notebook data and metadata, but the underlying data generation and accurate capture of experimental nuance still requires human input and verification, limiting full end-to-end automation.
Adoption barriersclaude-haiku-4-5-202510013/5Regulatory requirements (FDA 21 CFR Part 11, GLP, ISO compliance) mandate audit trails and human accountability for data integrity, creating oversight obligations that slow pure automation. Data privacy (HIPAA, institutional review boards) and instrument-specific validation add friction, though they do not legally require a human to manually enter every datum.
Adoption barriersclaude-sonnet-53/5Regulatory requirements (e.g., GLP, data integrity standards, reproducibility norms) mean records often need human verification and signatures, creating moderate friction against full automation.
Cost vs. human wageclaude-haiku-4-5-202510015/5Once integrated into a lab information system, AI-driven record maintenance (inference + automated logging) costs pennies per record versus several dollars for human manual data entry and cross-checking, yielding 10–100× cost savings at scale.
Cost vs. human wageclaude-sonnet-53/5ELN and AI-assisted documentation tools reduce time spent on record-keeping but require licensing, integration, and human oversight, making cost savings moderate rather than order-of-magnitude cheaper.
Technical feasibility todayclaude-haiku-4-5-202510014/5Laboratory data management platforms with AI-assisted structuring and cloud-based record systems are deployed in production across academic and pharmaceutical settings; OCR for handwritten notes, automated instrument integration, and compliance-aware data logging are mature. Error rates on routine entry are low, though complex or ambiguous entries still need human review.
Technical feasibility todayclaude-sonnet-53/5Electronic lab notebook (ELN) software with AI-assisted tagging, OCR, and data entry exists and is used in labs, but reliable automatic capture of accurate experimental records without human review is not yet standard practice.

Verify all financial, physical, and human resources assigned to research or development projects are used as planned.

61

CI 34–87 · exposure 58 · augmentation 63 · importance 3.2/5 · click for rater detail

Panel ratings per dimension for this task
DimensionModelRatingRationale
Sector adoption velocityclaude-haiku-4-5-202510014/5Large research institutions and biotech firms are actively deploying AI-powered compliance and resource-tracking systems; adoption is particularly rapid in well-funded sectors with mature IT infrastructure and pressure to reduce administrative overhead.
Sector adoption velocityclaude-sonnet-52/5Academic and scientific research administration is generally slow to adopt AI-driven financial/resource oversight tools compared to fast-moving finance or tech sectors.
Augmentation potentialclaude-haiku-4-5-202510014/5AI excels at flagging resource discrepancies and generating audit reports, enabling researchers and project managers to focus on root-cause investigation and corrective action rather than manual data collection and reconciliation.
Augmentation potentialclaude-sonnet-53/5AI can help track resource usage, flag anomalies, and generate reports that assist a human in verifying compliance, improving efficiency without replacing the oversight function.
Task automatabilityclaude-haiku-4-5-202510015/5AI systems can fully automate verification of financial, physical, and human resource allocation against plans by analyzing project documentation, timesheets, purchase orders, and budgets—a data-matching task with clear success criteria and >50% time savings achievable with current tools.
Task automatabilityclaude-sonnet-52/5This task involves cross-checking budgets, resource allocations, and personnel usage against project plans, which requires judgment about deviations and context AI can only partially support; full verification with accountability is not yet reliably automatable end-to-end.
Adoption barriersclaude-haiku-4-5-202510012/5While some institutional policies require human sign-off on budget compliance, there are no legal or licensing barriers preventing AI verification; the main friction is organizational preference for human oversight and integration with legacy ERP systems.
Adoption barriersclaude-sonnet-53/5While no explicit licensure is required, institutional accountability, grant compliance rules, and audit responsibility create moderate organizational friction against full automation.
Cost vs. human wageclaude-haiku-4-5-202510015/5AI-driven resource verification (via automated auditing tools and RPA) costs orders of magnitude less than hiring full-time research administrators or finance staff to manually reconcile project budgets, time sheets, and procurement records.
Cost vs. human wageclaude-sonnet-53/5AI-assisted tracking tools can cut some manual reconciliation costs, but human oversight and judgment remain necessary, keeping costs roughly comparable to a human doing this verification work.
Technical feasibility todayclaude-haiku-4-5-202510014/5Deployed financial audit and project management software with AI-powered anomaly detection already performs similar verification tasks in production, though integration into lab-specific resource tracking workflows is still maturing in some organizations.
Technical feasibility todayclaude-sonnet-52/5Some financial and project management software includes automated tracking and flagging dashboards, but true verification requiring domain judgment and sign-off is not handled reliably by deployed products in scientific research settings.

Prepare or review reports, manuscripts, or meeting presentations.

44

CI 30–59 · exposure 42 · augmentation 88 · importance 4.2/5 · click for rater detail

Panel ratings per dimension for this task
DimensionModelRatingRationale
Sector adoption velocityclaude-haiku-4-5-202510012/5While academic and biotech organizations experiment with AI writing tools, adoption remains cautious and largely confined to preliminary drafting. Production adoption is slow because reviewers, journals, and institutions still demand substantial human expert involvement and accountability.
Sector adoption velocityclaude-sonnet-53/5Academic and biotech research settings show growing but uneven adoption of AI writing tools, with usage more common for drafting than for official reporting or presentations.
Augmentation potentialclaude-haiku-4-5-202510014/5AI meaningfully assists by accelerating initial drafts, reorganizing sections, and catching obvious clarity issues, allowing the biologist to focus on data interpretation and critical revision. When used as a drafting partner rather than replacement, productivity gains are substantial while human expertise remains central.
Augmentation potentialclaude-sonnet-55/5AI substantially speeds up drafting, summarizing data, generating figures/slides, and improving clarity, while the scientist retains responsibility for accuracy and interpretation.
Task automatabilityclaude-haiku-4-5-202510012/5Current AI can assist with drafting and structure, but reviews of scientific reports and manuscripts require deep domain knowledge, interpretation of experimental data, and critical evaluation that remain difficult to automate end-to-end. AI typically produces 20–40% time savings with substantial human oversight, falling short of the 50% threshold.
Task automatabilityclaude-sonnet-53/5AI can draft manuscript sections, summarize data, and format presentations, but reviewing scientific accuracy and interpreting novel results still requires expert human judgment, limiting full end-to-end automation.
Adoption barriersclaude-haiku-4-5-202510014/5Scientific integrity, institutional review standards, and journal/conference submission policies create friction: manuscripts require expert authorship signatures, institutional accountability, and peer review. Liability for scientific errors and the expectation that senior scientists personally vouch for accuracy create significant adoption barriers.
Adoption barriersclaude-sonnet-52/5No formal licensing requirement for authorship, but journal policies, peer review norms, and authorship accountability standards create moderate friction against pure AI-generated reports.
Cost vs. human wageclaude-haiku-4-5-202510012/5AI inference costs are low, but the expert biologist must still spend substantial time reviewing, fact-checking, and revising AI-generated content, plus integration overhead. The all-in cost remains competitive with or slightly higher than having the scientist write directly.
Cost vs. human wageclaude-sonnet-54/5AI drafting and editing tools cost a small fraction of a scientist's hourly wage, though human review/oversight time still adds substantial cost relative to fully autonomous output.
Technical feasibility todayclaude-haiku-4-5-202510013/5Products like Claude, GPT-4, and specialized scientific writing tools can draft sections and provide feedback, but they produce material errors in technical accuracy, miss subtle methodological flaws, and require expert human review before publication or presentation. Deployment is limited to preliminary drafting stages.
Technical feasibility todayclaude-sonnet-53/5Tools like ChatGPT, Grammarly, and specialized writing assistants are widely used by scientists for drafting and editing, but review for scientific rigor and correctness remains a manual, error-prone process for AI.

Develop guidelines for procedures such as the management of viruses.

40

CI 25–55 · exposure 45 · augmentation 75 · importance 3.5/5 · click for rater detail

Panel ratings per dimension for this task
DimensionModelRatingRationale
Sector adoption velocityclaude-haiku-4-5-202510012/5Life-sciences organizations are still early in AI adoption for procedural documentation due to regulatory conservatism, institutional inertia, and the safety-critical nature of biosafety protocols. Pilots exist but production displacement remains minimal in this highly regulated space.
Sector adoption velocityclaude-sonnet-52/5Academic and biomedical research settings have been slower and more cautious in adopting AI for safety-critical documentation compared to fast-moving digital/finance sectors.
Augmentation potentialclaude-haiku-4-5-202510014/5AI assistants already help biologists dramatically by searching and summarizing vast literature, flagging regulatory updates, and generating structured first drafts that the expert then refines. This substantially accelerates guideline development while preserving expert judgment on safety and compliance.
Augmentation potentialclaude-sonnet-54/5AI can meaningfully speed up literature review, drafting, and formatting of procedural guidelines, letting scientists focus on technical accuracy and compliance review.
Task automatabilityclaude-haiku-4-5-202510014/5AI can now generate draft procedural guidelines by synthesizing published literature, regulatory frameworks, and best practices at scale, substantially reducing time-to-first-draft. However, final validation against institutional biosafety policies and regulatory compliance still requires expert judgment, limiting full end-to-end automation.
Task automatabilityclaude-sonnet-52/5Drafting guideline documents requires synthesizing scientific literature, safety standards, and institutional context, which AI can assist with but not reliably originate end-to-end at expert quality without heavy human oversight.rp Current AI can produce drafts but not authoritative guidelines independently.
Adoption barriersclaude-haiku-4-5-202510014/5Regulatory bodies and institutional biosafety committees typically require a qualified human biologist to author, review, and sign off on viral handling procedures; liability and legal certification are tightly coupled to credentialed personnel. Automated guidelines alone cannot satisfy these mandates.
Adoption barriersclaude-sonnet-54/5Guidelines involving virus management often require biosafety officer sign-off, regulatory compliance (e.g., CDC/NIH biosafety standards), and institutional accountability, creating strong barriers to full automation.
Cost vs. human wageclaude-haiku-4-5-202510014/5Using AI for literature synthesis and initial drafting costs far less than a biologist's fully-loaded wage per task, with incremental human expert review to validate and finalize. The AI handles the high-effort synthesis phase, leaving humans the judgment-intensive review.
Cost vs. human wageclaude-sonnet-52/5While AI drafting is cheap, the need for expert scientific review, institutional approval, and liability checking means overall cost savings versus a qualified biologist are modest, not dramatic.
Technical feasibility todayclaude-haiku-4-5-202510013/5Large language models and biomedical AI tools can generate reasonable guideline drafts and flag relevant literature, but production deployment remains limited because institutional review and legal/safety sign-off are still required. No mainstream product yet fully owns this task from inception to deployment.
Technical feasibility todayclaude-sonnet-52/5No deployed product autonomously creates biosafety or virus-management guidelines used in production labs; LLMs are used as drafting aids but outputs require substantial expert review and validation.

Write grant applications to obtain funding.

37

CI 29–45 · exposure 33 · augmentation 75 · importance 4.4/5 · click for rater detail

Panel ratings per dimension for this task
DimensionModelRatingRationale
Sector adoption velocityclaude-haiku-4-5-202510012/5Adoption remains limited; researchers use AI for drafting help (pilots, early adoption), but full automation is not yet standard practice in academic or research institutions due to quality concerns and institutional resistance.
Sector adoption velocityclaude-sonnet-53/5Academic science is a middling-adoption sector; AI writing assistants are increasingly used for drafts but grant offices and PIs remain cautious, with production-scale deployment still uneven.
Augmentation potentialclaude-haiku-4-5-202510014/5AI provides strong productivity gains as an assistant: generating outlines, drafting background sections, identifying literature, refining language, and flagging narrative gaps, allowing biologists to focus on novel science and strategic positioning of their research.
Augmentation potentialclaude-sonnet-54/5AI substantially helps with drafting introductions, formatting, editing for clarity, and summarizing literature, meaningfully speeding up the writing process while the scientist retains control over content and strategy.
Task automatabilityclaude-haiku-4-5-202510012/5While AI can draft sections (background, literature review, budget justification) with modest time savings, grant applications require novel intellectual synthesis, addressing specific funder priorities, and deep domain justification that demand human expertise. Current systems cannot reliably achieve 50% time savings at equal quality end-to-end.
Task automatabilityclaude-sonnet-53/5AI can draft grant sections, background, and boilerplate text given inputs, but crafting a competitive novel research plan, aligning with reviewer expectations, and integrating preliminary data requires substantial human expertise and iteration, limiting full end-to-end automation.
Adoption barriersclaude-haiku-4-5-202510014/5Intellectual property, attribution, and funder expectations create friction: many institutions and funding agencies have policies on what can be AI-generated, require human certification of grant content, and expect authentic researcher voice, limiting straightforward substitution.
Adoption barriersclaude-sonnet-53/5No licensing requirement, but funding agencies expect PI authorship, scientific accountability, and originality, creating institutional and reputational friction against full automation.
Cost vs. human wageclaude-haiku-4-5-202510013/5AI inference and integration costs are low compared to a biologist's hourly rate, but human review, revision, and quality control are still required, making the net cost per fundable application roughly comparable to human-only effort.
Cost vs. human wageclaude-sonnet-53/5AI drafting tools are cheap relative to PI/postdoc time spent writing, but the residual human effort for scientific content, strategy, and revision keeps overall cost savings moderate rather than order-of-magnitude.
Technical feasibility todayclaude-haiku-4-5-202510012/5AI tools exist for writing assistance and outlining, but no production system reliably writes full, fundable grant applications from scratch. Success rates would be material because funders expect novel scientific contribution tailored to their review criteria.
Technical feasibility todayclaude-sonnet-52/5LLM-based writing tools are used to draft portions of grants, but no deployed product reliably produces complete, fundable grant applications without heavy scientist involvement and editing.

Compile and analyze molecular or cellular experimental data and adjust experimental designs as necessary.

35

CI 32–38 · exposure 25 · augmentation 75 · importance 4.1/5 · click for rater detail

Panel ratings per dimension for this task
DimensionModelRatingRationale
Sector adoption velocityclaude-haiku-4-5-202510013/5Biotech and pharmaceutical R&D have adopted computational tools for data management and basic statistical analysis at scale, but autonomous experimental design adjustment remains in pilot/prototype phase; most labs still rely on human expertise for design decisions.
Sector adoption velocityclaude-sonnet-53/5Life sciences research increasingly uses AI/ML tools for data analysis (e.g., in genomics, proteomics), representing moderate and growing adoption, though wet-lab experimental design remains largely human-driven.
Augmentation potentialclaude-haiku-4-5-202510014/5AI systems (statistical packages, machine learning for pattern detection in high-dimensional data, experimental design software) meaningfully augment biologists' productivity in compiling, visualizing, and identifying trends in experimental data, though the human typically retains responsibility for design validation and hypothesis-driven adjustments.
Augmentation potentialclaude-sonnet-54/5AI significantly augments this task by automating statistical analysis, pattern detection, and literature-informed suggestions for experimental adjustments, substantially boosting researcher productivity while humans retain design authority.
Task automatabilityclaude-haiku-4-5-202510012/5While AI can assist with data compilation and statistical analysis of molecular/cellular datasets, the task requires judgment-driven experimental design adjustments that depend on tacit domain knowledge, unexpected results interpretation, and iterative hypothesis refinement—none of which current systems can reliably perform end-to-end without extensive human oversight.
Task automatabilityclaude-sonnet-52/5AI can assist with data compilation, statistical analysis, and generating hypotheses for redesign, but the physical experimentation, judgment about biological plausibility, and iterative wet-lab decision-making require human expertise not fully replaceable today.
Adoption barriersclaude-haiku-4-5-202510013/5Lab settings typically require institutional approval (IRBs, lab safety protocols) and professional accountability for experimental validity; however, there is no absolute licensing requirement preventing computational systems from assisting in data analysis and design iteration, so barriers are moderate rather than hard.
Adoption barriersclaude-sonnet-52/5No formal licensing prevents AI from assisting with data analysis, but institutional and publication norms require human scientists to be accountable for experimental design and interpretation, creating moderate organizational friction.
Cost vs. human wageclaude-haiku-4-5-202510012/5Data analysis tools are inexpensive, but the high-touch nature of experimental design adjustment and the need for expert human review make the all-in cost comparable to or potentially higher than employing a junior analyst for this hybrid task.
Cost vs. human wageclaude-sonnet-52/5While AI tools can cheaply crunch large datasets, the overall task still requires expensive PhD-level scientific judgment for experimental redesign, keeping the all-in cost of full automation comparable to or higher than human cost.
Technical feasibility todayclaude-haiku-4-5-202510012/5Deployed products exist for data visualization and statistical analysis (e.g., standard bioinformatics pipelines, R/Python libraries), but no production system reliably handles the interpretive loop of analyzing experimental data and autonomously adjusting experimental designs with comparable quality to trained biologists.
Technical feasibility todayclaude-sonnet-52/5Products like bioinformatics pipelines and AI copilots (e.g., for genomics data analysis) exist but are narrow, task-specific, and require significant human oversight to interpret results and redesign experiments reliably.

Evaluate new technologies to enhance or complement current research.

31

CI 25–38 · exposure 25 · augmentation 75 · importance 4.1/5 · click for rater detail

Panel ratings per dimension for this task
DimensionModelRatingRationale
Sector adoption velocityclaude-haiku-4-5-202510012/5Life sciences research remains relatively slow in adopting autonomous AI systems for decision-making tasks, partly due to regulatory scrutiny, institutional conservatism, and the high stakes of methodological choices. Adoption is primarily limited to assistive tools rather than autonomous systems.
Sector adoption velocityclaude-sonnet-53/5Academic and biotech research settings show growing but uneven adoption of AI tools for literature review, hypothesis generation, and technology scouting, still short of deep production-level integration.
Augmentation potentialclaude-haiku-4-5-202510014/5AI can meaningfully assist by rapidly summarizing technology papers, comparing specifications across competing platforms, identifying relevant publications, and highlighting technical trade-offs—leaving the final evaluation and decision to the expert biologist. This augmentation can substantially accelerate the evaluation process while maintaining human oversight.
Augmentation potentialclaude-sonnet-54/5AI tools (e.g., literature synthesis, patent/technology scanning, trend analysis) meaningfully speed up scanning and evaluating new methods, though the final judgment and decision-making stay human-led.
Task automatabilityclaude-haiku-4-5-202510012/5Evaluating new technologies requires nuanced scientific judgment, critical appraisal of peer-reviewed literature, and domain expertise to assess fit with research goals. While AI can summarize technology papers or compile feature comparisons, the core task of evaluating suitability, interpreting technical limitations, and making recommendations demands human expert judgment that current AI cannot reliably replicate end-to-end.
Task automatabilityclaude-sonnet-52/5This requires synthesizing scientific literature, understanding lab-specific context, and exercising judgment about experimental fit, which AI can support but not fully perform end-to-end today.
Adoption barriersclaude-haiku-4-5-202510014/5Research institutions and funding bodies typically require that technology evaluation and adoption decisions be made by qualified human scientists who are accountable for the rigor and validity of those choices. Professional standards and institutional governance create meaningful barriers to full automation.
Adoption barriersclaude-sonnet-52/5No licensing requirement, but institutional trust, funding decisions, and technical credibility mean human expert judgment remains central with only moderate friction to AI involvement.
Cost vs. human wageclaude-haiku-4-5-202510012/5The cost of AI assistance (literature analysis, comparison tools) plus oversight by a molecular biologist approximates or exceeds the cost of a biologist conducting the evaluation directly, particularly given the need for expert verification of recommendations.
Cost vs. human wageclaude-sonnet-52/5AI can cheaply summarize papers and technology reports, but the actual evaluation still requires expert scientist time for validation, feasibility testing, and integration decisions, keeping overall cost comparable to human-led effort.
Technical feasibility todayclaude-haiku-4-5-202510012/5No deployed products reliably perform comprehensive technology evaluation for research contexts. AI systems can assist with literature summarization or data extraction, but production systems do not independently evaluate novel biotech tools against research-specific criteria with the accuracy required for research decisions.
Technical feasibility todayclaude-sonnet-52/5AI literature-review and summarization tools exist and are used by researchers, but no deployed product independently evaluates and recommends new technologies for a specific research program reliably.

Design molecular or cellular laboratory experiments, oversee their execution, and interpret results.

29

CI 25–32 · exposure 25 · augmentation 75 · importance 4.5/5 · click for rater detail

Panel ratings per dimension for this task
DimensionModelRatingRationale
Sector adoption velocityclaude-haiku-4-5-202510012/5Biotech and pharmaceutical sectors are adopting AI for narrow tasks (high-throughput screening analysis, molecular modeling), but full experimental design and execution remain human-driven; adoption is in pilot phase for most labs rather than production displacement of the integrated task.
Sector adoption velocityclaude-sonnet-53/5Biotech and pharma sectors are increasingly using AI for data analysis and literature review, but full experimental design/oversight automation remains at pilot stage rather than broad production deployment.
Augmentation potentialclaude-haiku-4-5-202510014/5AI significantly augments productivity on this task—literature mining, protocol optimization, statistical analysis, and data interpretation are all enhanced by current tools—while biologists retain essential judgment about feasibility, safety, and contextual meaning of results.
Augmentation potentialclaude-sonnet-54/5AI substantially aids experimental design through literature synthesis, statistical planning, and data interpretation, significantly boosting researcher productivity while the scientist remains central to judgment calls and physical execution.
Task automatabilityclaude-haiku-4-5-202510012/5Current AI can assist with experimental design (literature review, hypothesis generation, statistical planning) and computational analysis of results, but the physical execution of experiments, troubleshooting unexpected outcomes, and contextual interpretation of complex biological data require ongoing human judgment and hands-on oversight that falls well short of 50% time savings at equal quality end-to-end.
Task automatabilityclaude-sonnet-52/5AI can assist with hypothesis generation, protocol drafting, and literature review, but designing rigorous experiments and overseeing physical lab execution requires hands-on judgment, adaptive troubleshooting, and physical manipulation that current AI cannot perform end-to-end.
Adoption barriersclaude-haiku-4-5-202510014/5Institutional requirements (Institutional Biosafety Committees, IRBs for certain studies, regulatory compliance in pharmaceutical/biotech settings) mandate qualified human oversight; liability for incorrect interpretation and potential biosafety failures creates strong legal and organizational barriers to autonomous automation.
Adoption barriersclaude-sonnet-53/5No formal licensing requirement blocks AI use, but institutional review, safety protocols, funding accountability, and scientific credibility norms create meaningful friction against full automation of experiment design and interpretation.
Cost vs. human wageclaude-haiku-4-5-202510012/5AI inference and analysis tools are relatively inexpensive, but the full task requires expert human supervision (senior scientist salary >$100k/year equivalent), and AI cannot yet meaningfully reduce total cost when integration, validation, and oversight are factored in.
Cost vs. human wageclaude-sonnet-52/5AI can cheaply assist with parts like data analysis or literature synthesis, but the overall task still requires expensive skilled labor for hands-on execution and oversight, keeping costs comparable to human labor.
Technical feasibility todayclaude-haiku-4-5-202510012/5While AI tools exist for literature synthesis, protocol suggestion, and data analysis, no deployed product reliably performs the full cycle of design, execution oversight, and result interpretation autonomously; production systems handle narrow subtasks (sequence alignment, statistical analysis) rather than the integrated experimental workflow.
Technical feasibility todayclaude-sonnet-52/5AI tools (e.g., literature-mining assistants, statistical analysis copilots) are used in research settings but no deployed product autonomously designs, executes, and interprets wet-lab experiments reliably at scale.

Conduct research on cell organization and function, including mechanisms of gene expression, cellular bioinformatics, cell signaling, or cell differentiation.

29

CI 25–32 · exposure 25 · augmentation 75 · importance 4.3/5 · click for rater detail

Panel ratings per dimension for this task
DimensionModelRatingRationale
Sector adoption velocityclaude-haiku-4-5-202510012/5Adoption of AI tools in biology is moderate and concentrated in data analysis and computational steps; wet-lab and experimental design phases remain human-driven. Most institutions use AI as a research aid rather than an autonomous agent, reflecting slower adoption in experimental science versus information-sector roles.
Sector adoption velocityclaude-sonnet-53/5Academic and biotech research settings are adopting AI tools (AlphaFold, bioinformatics pipelines) at a moderate pace, with pilots and specific tool integration common but full workflow automation rare.
Augmentation potentialclaude-haiku-4-5-202510014/5AI strongly augments research productivity through literature mining, sequence analysis, structure prediction, and data interpretation. Tools like AlphaFold and bioinformatics platforms significantly accelerate hypothesis generation and validation when used alongside human expertise and experimental work.
Augmentation potentialclaude-sonnet-54/5AI substantially assists literature review, sequence/data analysis, hypothesis generation, and experimental design, meaningfully boosting researcher productivity while humans retain control of experiments and interpretation.
Task automatabilityclaude-haiku-4-5-202510012/5Current AI can assist with literature mining, sequence analysis, and computational modeling, but experimental design, hypothesis formation, and wet-lab execution remain human-dependent. The task involves high-level research direction and creative problem-solving that cannot be fully automated to the 50% time-saving threshold.
Task automatabilityclaude-sonnet-52/5Core research requires wet-lab experimentation, hypothesis generation, and physical manipulation of biological materials that current AI cannot perform end-to-end; AI aids analysis but not the full research cycle.core
Adoption barriersclaude-haiku-4-5-202510014/5Regulatory oversight (IRB approval, biosafety), institutional credentialing, and liability requirements for novel findings mean only licensed researchers can publish and act on results. Scientific publication norms and peer review require human authorship and accountability, creating hard barriers to full substitution.
Adoption barriersclaude-sonnet-53/5No strict licensing requirement for research itself, but institutional review, funding accountability, and scientific reproducibility norms create moderate friction against full automation.
Cost vs. human wageclaude-haiku-4-5-202510012/5AI inference costs for analysis are low, but integration, wet-lab equipment, reagents, and required human supervision dominate the total cost. The human expert wage remains the primary cost driver for research-quality work in this domain.
Cost vs. human wageclaude-sonnet-52/5Wet-lab equipment, reagents, and human expertise dominate costs; AI reduces some computational analysis time but does not replace the bulk of research costs, so savings are partial not order-of-magnitude.
Technical feasibility todayclaude-haiku-4-5-202510012/5Deployed products exist for bioinformatics pipelines and data analysis, but no end-to-end AI system reliably conducts independent cell biology research. Deployed tools are narrow in scope (e.g., sequence alignment, image analysis) and require substantial human oversight and interpretation.
Technical feasibility todayclaude-sonnet-52/5Deployed AI tools exist for bioinformatics analysis (sequence alignment, expression analysis) but no product autonomously conducts cell biology research including wet-lab experimentation and interpretation.

Instruct undergraduate and graduate students within the areas of cellular or molecular biology.

28

CI 25–30 · exposure 25 · augmentation 75 · importance 4.2/5 · click for rater detail

Panel ratings per dimension for this task
DimensionModelRatingRationale
Sector adoption velocityclaude-haiku-4-5-202510012/5Despite digital transformation in higher ed, actual AI replacement of instructional tasks remains limited to pilot chatbot tutoring and lecture transcription. Most adoption is assistive (draft slides, student Q&A support), not substitution, and faculty resistance remains high in academic sectors.
Sector adoption velocityclaude-sonnet-52/5Higher education adopts AI tools slowly for core instruction due to academic norms, accreditation, and quality concerns, though tools for course prep are gaining traction.
Augmentation potentialclaude-haiku-4-5-202510014/5AI can meaningfully assist instructors by drafting lecture slides, generating practice problems, providing automated grading for objective assessments, and offering round-the-clock student tutoring support. These tools can substantially raise instructor productivity in preparation and student engagement while keeping the instructor in the loop for judgment and mentorship.
Augmentation potentialclaude-sonnet-54/5AI substantially aids instructors in preparing lecture materials, generating practice problems, explaining concepts, and answering student questions, enhancing teaching productivity.
Task automatabilityclaude-haiku-4-5-202510012/5While AI can generate lecture content and explain concepts, teaching requires real-time interaction, assessment of student understanding, adaptation to questions, and mentorship—tasks where current AI systems cannot reliably replace human instructors at scale. Lecture recording and content generation might save ~20–30% of preparation time, but the interactive teaching component remains largely manual.
Task automatabilityclaude-sonnet-52/5Teaching involves live interaction, mentorship, and adaptive responses to student needs that current AI cannot fully replicate end-to-end, though AI can help create materials.
Adoption barriersclaude-haiku-4-5-202510014/5Regulatory and institutional barriers are substantial: universities accredit and credential human instructors, students expect human mentorship and accountability, and labor agreements protect faculty roles. Legal liability for educational outcomes and the requirement that instructors hold graduate degrees and teach under institutional authority create hard adoption friction.
Adoption barriersclaude-sonnet-53/5No formal licensing requires a human instructor for all teaching, but accreditation standards, university policies, and expectations of faculty-student mentorship create moderate friction.
Cost vs. human wageclaude-haiku-4-5-202510012/5Current AI systems (content generation, tutoring bots) cost roughly $0.01–0.10 per student interaction, but a full course requires instructor oversight, feedback calibration, and credentialing that AI cannot provide. The all-in cost remains higher than a human instructor at typical academic wages when institutional requirements are met.
Cost vs. human wageclaude-sonnet-52/5AI can generate lecture content or quizzes cheaply, but human oversight, lab supervision, and mentorship remain necessary, keeping overall cost comparable to human instruction.
Technical feasibility todayclaude-haiku-4-5-202510012/5AI systems can draft lesson materials and answer student questions via chatbots, but no deployed product reliably performs full course instruction (grading nuance, Socratic dialogue, mentorship, accountability for student outcomes). Institutional resistance and accreditation requirements mean adoption is narrow and supplemental only.
Technical feasibility todayclaude-sonnet-52/5AI tutoring products exist for content delivery but no deployed system reliably substitutes for an instructor's full instructional role in specialized biology courses.

Monitor or operate specialized equipment, such as gas chromatographs and high pressure liquid chromatographs, electrophoresis units, thermocyclers, fluorescence activated cell sorters, and phosphorimagers.

28

CI 25–30 · exposure 25 · augmentation 50 · importance 3.8/5 · click for rater detail

Panel ratings per dimension for this task
DimensionModelRatingRationale
Sector adoption velocityclaude-haiku-4-5-202510012/5Academic and biotech labs are digitizing data management and analysis, but actual equipment operation remains largely manual by design and regulation. Adoption of AI for autonomous instrument operation is minimal; most labs are in the pilot or evaluation phase at best.
Sector adoption velocityclaude-sonnet-52/5Biotech and pharma sectors are investing in lab automation, but adoption is uneven and many academic and smaller labs still rely heavily on manual equipment operation.'
Augmentation potentialclaude-haiku-4-5-202510013/5AI can assist by analyzing real-time data streams, flagging anomalies, suggesting parameter adjustments, and automating post-run data processing and reporting. These capabilities improve technician productivity on monitoring and interpretation tasks, though the human remains essential for hands-on operation and problem-solving.
Augmentation potentialclaude-sonnet-53/5AI-assisted data interpretation, automated protocol optimization, and software controlling instrument parameters can meaningfully boost efficiency, but the human still directly runs and monitors the physical equipment.'
Task automatabilityclaude-haiku-4-5-202510012/5While AI can assist with data interpretation and analysis, the hands-on operation of specialized laboratory equipment—sample loading, parameter adjustment, maintenance, troubleshooting hardware issues—requires physical manipulation and real-time adaptive decision-making that current AI systems cannot perform end-to-end. Monitoring outputs is partially automatable, but overall task completion falls short of 50% time savings.
Task automatabilityclaude-sonnet-52/5Physical operation of lab instruments, sample loading, and troubleshooting requires manual dexterity and real-time judgment that current AI cannot perform; software can automate data analysis but not the hands-on equipment operation.'
Adoption barriersclaude-haiku-4-5-202510014/5Regulatory compliance (ISO 17025, GLP, GMP standards) and quality assurance requirements mandate human accountability and sign-off on results from specialized analytical equipment. Liability for incorrect operation or data integrity issues creates strong institutional friction against unsupervised AI automation.
Adoption barriersclaude-sonnet-53/5No licensing requirement per se, but safety protocols, equipment calibration, and quality control in research settings create organizational friction and liability concerns around unattended operation.'
Cost vs. human wageclaude-haiku-4-5-202510012/5AI solutions for equipment operation (if they existed in production form) would require expensive custom integration, continuous oversight, and human technician supervision. The loaded cost of a skilled molecular biologist or lab technician ($50–80K/year) remains lower than full automation deployment for specialized equipment.
Cost vs. human wageclaude-sonnet-52/5Automated lab robotics systems are expensive to acquire and integrate, often costing more than a technician's wages for low-to-moderate throughput operations, though savings occur at high scale.'
Technical feasibility todayclaude-haiku-4-5-202510012/5No deployed AI system reliably operates these instruments end-to-end in production labs today. Some software exists for data analysis and reporting post-run, but autonomous equipment operation and troubleshooting remain research-stage; most labs still require trained humans at the instrument.
Technical feasibility todayclaude-sonnet-52/5Lab automation platforms and robotic liquid handlers exist for narrow, standardized workflows, but general monitoring/operating of diverse specialized equipment like FACS or phosphorimagers still requires trained personnel in most labs.'

Participate in all levels of bioproduct development, including proposing new products, performing market analyses, designing and performing experiments, and collaborating with operations and quality control teams during product launches.

28

CI 25–30 · exposure 25 · augmentation 75 · importance 3.1/5 · click for rater detail

Panel ratings per dimension for this task
DimensionModelRatingRationale
Sector adoption velocityclaude-haiku-4-5-202510012/5Biotech and pharmaceutical sectors are early in AI adoption for this type of integrated product development role. While AI-assisted drug discovery and data analysis are growing, autonomous or near-autonomous participation in full bioproduct development cycles remains experimental; human scientists remain central to decision-making.
Sector adoption velocityclaude-sonnet-52/5Biotech/pharma R&D is adopting AI for specific analytic tasks (e.g., literature mining, market forecasting) but lab-based product development and quality control processes remain slow to digitize compared to purely information-based sectors.
Augmentation potentialclaude-haiku-4-5-202510014/5AI can meaningfully assist with market trend analysis, literature mining, experimental design suggestions, and data interpretation, helping biologists work faster and more comprehensively. However, the human scientist remains the critical decision-maker for feasibility, safety, and strategic direction, making this a strong augmentation scenario.
Augmentation potentialclaude-sonnet-54/5AI can meaningfully assist with market analysis, experimental design ideation, literature review, and report drafting, improving efficiency across several sub-tasks even though humans remain central to execution and decision-making.
Task automatabilityclaude-haiku-4-5-202510012/5While AI can assist with literature review, experimental design documentation, and market analysis, the task requires end-to-end judgment about biological feasibility, regulatory strategy, and cross-functional collaboration that current systems cannot perform autonomously. Proposing viable new products and navigating the full development lifecycle remain heavily human-dependent.
Task automatabilityclaude-sonnet-52/5This task spans strategic ideation, hands-on wet-lab experimentation, and cross-team collaboration during launches, most of which require physical lab work, domain judgment, and interpersonal coordination that current AI cannot execute end-to-end.15% time-savings are plausible on analysis/writing sub-tasks but not the whole task.
Adoption barriersclaude-haiku-4-5-202510014/5Regulatory requirements (FDA, GMP compliance), liability for product safety and efficacy claims, and the need for human scientific judgment and accountability create substantial barriers. A licensed scientist must ultimately sign off on bioproduct safety and quality; automation cannot replace that legal and professional responsibility.
Adoption barriersclaude-sonnet-53/5While no explicit licensure is required for biologists doing this work, regulatory compliance (e.g., quality control, GMP) and organizational accountability for product launches impose meaningful oversight requirements that resist full automation.
Cost vs. human wageclaude-haiku-4-5-202510012/5AI inference for market analysis or document review is cheap, but integration into a molecular biologist's workflow (oversight, validation, rework) and the cost of errors in experimental design or product strategy mean total delivered cost remains comparable to or exceeds direct human labor for this complex task.
Cost vs. human wageclaude-sonnet-52/5AI can cheaply assist with market research or data analysis components, but the wet-lab experimentation and interdisciplinary coordination still require highly paid specialized human labor and equipment, keeping overall cost comparable or AI-supplemented rather than substitutive.
Technical feasibility todayclaude-haiku-4-5-202510012/5AI tools (LLMs, data analysis platforms) exist for components like market analysis and literature synthesis, but no deployed product performs the full integration of product proposal, experimental design, execution oversight, and cross-team collaboration at the quality required for bioproduct development. Production deployment in this domain remains rare.
Technical feasibility todayclaude-sonnet-52/5AI tools exist for literature synthesis, market analysis drafting, and experimental design suggestions, but no deployed product autonomously runs bioproduct development lifecycles including physical experimentation and cross-functional launch coordination.

Design databases, such as mutagenesis libraries.

28

CI 25–30 · exposure 25 · augmentation 63 · importance 2.5/5 · click for rater detail

Panel ratings per dimension for this task
DimensionModelRatingRationale
Sector adoption velocityclaude-haiku-4-5-202510012/5Adoption of AI for database design in molecular biology is nascent; most labs still rely on manual or semi-manual approaches. Pilot projects exist, but production-scale displacement is not evident in genomics/cell biology sectors.
Sector adoption velocityclaude-sonnet-52/5Biotech and academic research sectors are only beginning to adopt AI-assisted design tools; broad production-scale adoption for library design specifically remains limited.
Augmentation potentialclaude-haiku-4-5-202510013/5AI can meaningfully assist by generating candidate database schemas, suggesting mutation designs based on sequence analysis, and automating routine data organization, allowing biologists to focus on validation and experimental logic. However, the augmentation is partial, not transformative.
Augmentation potentialclaude-sonnet-54/5AI can meaningfully assist by generating candidate mutation sets, predicting effects, and optimizing library diversity, substantially speeding up the design process while the scientist retains oversight.
Task automatabilityclaude-haiku-4-5-202510012/5While AI can assist in data organization and schema design, designing mutagenesis libraries requires domain expertise to define biological parameters, mutation strategies, and experimental constraints that vary significantly by research goal. End-to-end automation would still require substantial human oversight and iterative refinement.
Task automatabilityclaude-sonnet-52/5Designing a mutagenesis library requires domain expertise, experimental context, and iterative wet-lab feasibility judgments that AI cannot fully replace, though it can assist with sequence design and combinatorics calculations.'
Adoption barriersclaude-haiku-4-5-202510014/5Regulatory oversight (IRB/biosafety review), institutional best practices for mutagenesis, and the need for expert biologists to sign off on library designs create meaningful barriers to full automation. Liability for flawed libraries also limits autonomous deployment.
Adoption barriersclaude-sonnet-53/5No strict licensing requirement, but scientific rigor, reproducibility standards, and grant/publication accountability create moderate organizational and quality-control friction against full automation.
Cost vs. human wageclaude-haiku-4-5-202510012/5AI infrastructure (LLMs, database tools) is relatively cheap, but the task still demands expert biologist time to validate designs, specify requirements, and integrate outputs with laboratory workflows. Total automation cost-benefit remains unfavorable compared to traditional expert design.
Cost vs. human wageclaude-sonnet-52/5AI tools can reduce some design time but still require significant expert review and integration with lab-specific constraints, so overall cost savings versus a trained scientist are modest.
Technical feasibility todayclaude-haiku-4-5-202510012/5No deployed products reliably design mutagenesis libraries end-to-end. AI tools can help with data structuring and query optimization, but the biological logic, validation of mutation sets, and alignment with experimental protocols remain primarily human-driven tasks in practice.
Technical feasibility todayclaude-sonnet-52/5Some computational tools and AI-assisted design platforms exist for library design (e.g., codon optimization, variant selection) but no mature product autonomously designs full mutagenesis libraries in production without expert oversight.

Conduct applied research aimed at improvements in areas such as disease testing, crop quality, pharmaceuticals, and the harnessing of microbes to recycle waste.

27

CI 21–32 · exposure 25 · augmentation 63 · importance 3.5/5 · click for rater detail

Panel ratings per dimension for this task
DimensionModelRatingRationale
Sector adoption velocityclaude-haiku-4-5-202510012/5While biotech and pharmaceutical firms are adopting AI for data analysis and drug discovery screening, adoption of AI-assisted applied research remains in pilot stages; most wet-lab work and experimental iteration still rely on human researchers, with AI playing a supporting rather than autonomous role.
Sector adoption velocityclaude-sonnet-53/5Biotech and pharma sectors are adopting AI tools for R&D (drug discovery, protein modeling) at a moderate-to-fast pace, with many pilots and growing production use in computational biology.
Augmentation potentialclaude-haiku-4-5-202510013/5AI tools meaningfully assist with experimental design optimization, statistical analysis of results, and literature synthesis, but the high barrier to full autonomy means augmentation is limited to specific sub-tasks rather than transforming the entire workflow of applied research.
Augmentation potentialclaude-sonnet-54/5AI substantially augments applied research through literature synthesis, experimental design suggestions, data analysis, and predictive modeling (e.g., protein folding, genomics), significantly boosting researcher productivity.
Task automatabilityclaude-haiku-4-5-202510012/5While AI can assist with literature review, data analysis, and hypothesis generation, the core experimental design, hands-on laboratory work, troubleshooting unexpected results, and creative problem-solving in applied research require significant human judgment and wet-lab execution that current AI cannot fully automate end-to-end.
Task automatabilityclaude-sonnet-52/5Applied research spans hypothesis generation, wet-lab experimentation, and physical validation that AI cannot yet perform end-to-end; AI can assist literature review, data analysis and hypothesis generation but the bulk of hands-on experimental work remains human-executed.19
Adoption barriersclaude-haiku-4-5-202510014/5Regulatory oversight (FDA, EPA, institutional biosafety committees), intellectual property concerns, liability for failed experiments or contaminated samples, and the requirement for credentialed scientists to certify experimental protocols and results create substantial legal and organizational barriers to full automation.
Adoption barriersclaude-sonnet-53/5No licensing requirement blocks AI-assisted research per se, but regulatory oversight (FDA, biosafety, IRB) and liability for pharmaceutical/agricultural claims impose meaningful friction on full automation.
Cost vs. human wageclaude-haiku-4-5-202510011/5The cost of running applied biology research (equipment, reagents, samples, regulatory compliance) far exceeds inference costs, and AI currently handles only peripheral tasks; the human researcher remains essential and expensive relative to AI assistance provided.
Cost vs. human wageclaude-sonnet-52/5Specialized AI tools are cheap for narrow computational subtasks, but overall research still requires costly human expertise, equipment, and wet-lab labor, keeping blended cost comparable to human costs.
Technical feasibility todayclaude-haiku-4-5-202510012/5AI tools exist for parts of the workflow (sequence analysis, literature mining, predictive modeling) but no deployed system can conduct applied biological research autonomously; the task requires iterative experimentation, physical manipulation of samples, and adaptive decision-making that remains beyond current AI capabilities.
Technical feasibility todayclaude-sonnet-52/5AI tools (e.g., AlphaFold, lab copilots) support specific subtasks like protein structure prediction or data analysis in production, but no deployed system performs full applied biological research reliably.

Develop assays that monitor cell characteristics.

23

CI 16–30 · exposure 20 · augmentation 63 · importance 3.5/5 · click for rater detail

Panel ratings per dimension for this task
DimensionModelRatingRationale
Sector adoption velocityclaude-haiku-4-5-202510012/5Biotech and pharmaceutical sectors are cautiously adopting AI for supporting assay work (data analysis, statistical interpretation), but core assay development remains deeply traditional and expert-driven. Production adoption of AI-designed assays is rare; most organizations still rely on human scientists for validation and regulatory acceptance.
Sector adoption velocityclaude-sonnet-52/5Biotech and pharma R&D are adopting AI tools for design assistance and data analysis, but wet-lab assay development itself sees slow, uneven uptake of autonomous AI methods.
Augmentation potentialclaude-haiku-4-5-202510013/5AI can meaningfully assist assay development through literature mining, statistical analysis of results, and optimization suggestions, raising efficiency in parts of the workflow. However, augmentation is limited by the need for human expertise in experimental design choices and the interpretability demands of validation; AI remains a supporting tool rather than a transformative one.
Augmentation potentialclaude-sonnet-54/5AI can meaningfully assist by suggesting assay designs, analyzing prior literature, predicting molecular interactions, and processing resulting data, significantly boosting researcher productivity while humans execute and validate experiments.
Task automatabilityclaude-haiku-4-5-202510012/5Assay development requires expert design judgment, validation protocols, and iterative troubleshooting that AI cannot fully automate. While AI can assist with literature review, statistical analysis, and protocol optimization, the core task of designing novel assays to monitor specific cell characteristics demands hands-on experimental insight and decision-making that remains beyond current end-to-end automation.
Task automatabilityclaude-sonnet-52/5Assay development requires iterative wet-lab experimentation, hypothesis generation, and physical troubleshooting that current AI cannot execute end-to-end; AI can assist in design but not perform the task itself.
Adoption barriersclaude-haiku-4-5-202510014/5Assay development is regulated by quality standards (GxP compliance, ISO standards), institutional biosafety committees, and often FDA oversight depending on application. Additionally, liability for assay validity and performance falls on human scientists who must sign off on protocols, creating strong legal barriers to full automation.
Adoption barriersclaude-sonnet-53/5No licensing requirement mandates a human specifically, but institutional review, reproducibility standards, and physical lab access create practical friction against full automation.
Cost vs. human wageclaude-haiku-4-5-202510012/5AI-assisted tools for assay design and analysis are still relatively expensive to deploy and require significant specialized infrastructure and expertise to interpret results. The computational and human oversight costs do not yet undercut the loaded cost of a molecular biologist performing this task, especially given quality and liability concerns.
Cost vs. human wageclaude-sonnet-51/5AI cannot replace the physical lab work, reagents, and hands-on experimentation needed, so the human scientist remains necessary and cost comparison favors humans doing the actual assay development.
Technical feasibility todayclaude-haiku-4-5-202510012/5No deployed AI systems reliably design and validate assays from first principles in production environments. While machine learning aids in data interpretation and some workflow steps, actual assay development requires experimental execution, failure analysis, and refinement cycles that current tools cannot independently perform at acceptable error rates.
Technical feasibility todayclaude-sonnet-51/5No deployed product autonomously develops and validates novel cell-based assays in production; this remains a research-stage capability requiring human scientists in the lab.

Evaluate new supplies and equipment to ensure operability in specific laboratory settings.

23

CI 16–30 · exposure 20 · augmentation 50 · importance 3.0/5 · click for rater detail

Panel ratings per dimension for this task
DimensionModelRatingRationale
Sector adoption velocityclaude-haiku-4-5-202510012/5Laboratory biologists work in regulated, high-stakes environments where adoption of fully automated equipment evaluation is slow; most organizations still rely on domain experts and vendor support for validation.
Sector adoption velocityclaude-sonnet-52/5Lab science and biology research sectors show moderate AI adoption for data analysis but lag in physical lab operations and equipment management.
Augmentation potentialclaude-haiku-4-5-202510013/5AI can meaningfully assist by summarizing equipment specs, cross-referencing compatibility databases, and flagging potential issues, helping the biologist focus on hands-on validation and contextual judgment.
Augmentation potentialclaude-sonnet-53/5AI can help research specifications, compare vendor equipment options, and summarize compatibility documentation, but cannot perform physical operability testing itself.
Task automatabilityclaude-haiku-4-5-202510012/5While AI can assist with spec review and comparison against predefined criteria, evaluating operability in specific laboratory settings requires hands-on testing, contextual judgment about equipment compatibility with existing systems, and validation of performance—tasks that largely exceed current automation capabilities.
Task automatabilityclaude-sonnet-52/5This involves hands-on physical testing of lab equipment compatibility with existing setups, which AI cannot physically perform; AI can only assist with research on specs beforehand.'
Adoption barriersclaude-haiku-4-5-202510014/5Laboratory safety regulations, equipment certification requirements, and institutional liability for faulty equipment validation create substantial barriers; a qualified scientist must typically validate operability and sign off on suitability decisions.
Adoption barriersclaude-sonnet-53/5No licensing barrier specifically, but physical presence, hands-on testing, and institutional procurement processes create practical friction against remote automation.
Cost vs. human wageclaude-haiku-4-5-202510012/5AI-assisted spec review might reduce time on document analysis, but the cost of oversight, validation, and human expertise needed to ensure correct evaluation remains high relative to the direct human labor saved.
Cost vs. human wageclaude-sonnet-51/5AI cannot substitute for the physical testing and hands-on verification required, so no meaningful cost comparison for automation exists yet.
Technical feasibility todayclaude-haiku-4-5-202510012/5No deployed products reliably perform end-to-end equipment evaluation in real laboratory contexts; AI can draft comparison matrices or flag spec mismatches, but actual operability verification still requires human validation and physical testing.
Technical feasibility todayclaude-sonnet-51/5No deployed product performs physical equipment evaluation and operability testing in labs; this remains a hands-on human task.

Direct, coordinate, organize, or prioritize biological laboratory activities.

23

CI 20–25 · exposure 16 · augmentation 50 · importance 4.2/5 · click for rater detail

Panel ratings per dimension for this task
DimensionModelRatingRationale
Sector adoption velocityclaude-haiku-4-5-202510012/5Life sciences research remains highly tradition-bound with slow digital transformation; while some labs use scheduling software, AI-driven laboratory management and coordination remain pilot-stage with minimal production adoption in mainstream academic or clinical settings.
Sector adoption velocityclaude-sonnet-52/5Academic and biotech labs adopt digital lab management and scheduling tools slowly, with AI-driven coordination still rare and pilots limited.
Augmentation potentialclaude-haiku-4-5-202510013/5AI can assist with workflow optimization, resource allocation suggestions, and data aggregation for prioritization decisions, but the human researcher must retain judgment authority over experimental direction and team coordination, providing moderate productivity gains.
Augmentation potentialclaude-sonnet-53/5AI can assist with scheduling, resource tracking, inventory management, and prioritization suggestions, meaningfully aiding a human lab director without replacing their judgment.
Task automatabilityclaude-haiku-4-5-202510012/5While AI can assist with scheduling and resource planning, directing and coordinating laboratory activities requires real-time human judgment, team management, problem-solving during failures, and adaptive prioritization based on experimental outcomes—tasks that remain primarily human-dependent today.
Task automatabilityclaude-sonnet-52/5Directing, coordinating, and prioritizing lab activities requires real-time judgment about experimental context, personnel, equipment, and shifting priorities that current AI cannot handle end-to-end.dejando
Adoption barriersclaude-haiku-4-5-202510014/5Laboratory management involves legal responsibility for safety compliance, regulatory adherence (OSHA, institutional biosafety), personnel supervision, and institutional accountability—typically requiring a credentialed human director or principal investigator to legally oversee operations.
Adoption barriersclaude-sonnet-53/5No licensing requirement specifically for this coordination task, but organizational accountability, safety protocols, and human oversight of lab operations create meaningful friction against full automation.
Cost vs. human wageclaude-haiku-4-5-202510012/5AI tools for scheduling and workflow optimization have low per-task cost, but coordinating biological laboratories at full scope still requires human oversight, training, and decision authority, making total substitution cost-prohibitive relative to senior staff wages.
Cost vs. human wageclaude-sonnet-52/5AI scheduling/project tools are cheap but cannot substitute for the managerial judgment and accountability required, so effective automation cost is not favorable relative to a human supervisor's output.
Technical feasibility todayclaude-haiku-4-5-202510011/5No deployed product reliably performs end-to-end laboratory direction and coordination; this requires contextual understanding of experimental priorities, team dynamics, equipment status, and adaptive decision-making that exceeds current AI capabilities in production settings.
Technical feasibility todayclaude-sonnet-51/5No deployed product autonomously manages and prioritizes biology lab operations; this remains a human management function with only ancillary scheduling tools available.

Confer with vendors to evaluate new equipment or reagents or to discuss the customization of product lines to meet user requirements.

21

CI 7–35 · exposure 13 · augmentation 50 · importance 2.6/5 · click for rater detail

Panel ratings per dimension for this task
DimensionModelRatingRationale
Sector adoption velocityclaude-haiku-4-5-202510012/5While research institutions are adopting digital procurement and AI-assisted documentation tools, the core task of conferring with vendors to evaluate customization remains a high-touch, human-driven activity; adoption of full vendor negotiation automation is minimal.
Sector adoption velocityclaude-sonnet-52/5Life sciences R&D adopts AI unevenly and mostly for data analysis or literature review; vendor negotiation and procurement processes remain largely manual and slow to digitize.
Augmentation potentialclaude-haiku-4-5-202510013/5AI can helpfully draft communication templates, summarize vendor specifications, track product comparisons, and organize requirements, assisting scientists in preparing for and documenting vendor interactions without replacing the human conferral.
Augmentation potentialclaude-sonnet-53/5AI tools can help scientists prepare specifications, compare vendor claims, and draft correspondence, improving efficiency without replacing the interpersonal negotiation process.
Task automatabilityclaude-haiku-4-5-202510011/5This task requires interpersonal negotiation, vendor relationship management, and domain-specific judgment about product customization needs. Current AI cannot autonomously conduct meaningful vendor conferences or make binding technical commitments.
Task automatabilityclaude-sonnet-52/5This involves relationship-based negotiation, technical evaluation, and back-and-forth judgment about lab-specific needs that current AI cannot conduct end-to-end, though AI can help draft communications or research vendor options.rag_1
Adoption barriersclaude-haiku-4-5-202510014/5Strong barriers exist: procurement and equipment decisions typically require explicit human sign-off by authorized personnel, organizational policies often mandate direct vendor contact by subject-matter experts, and liability for product selection failures typically rests with the scientist or institution.
Adoption barriersclaude-sonnet-52/5No formal licensing barrier, but organizational trust, vendor relationships, and the need for domain expertise to judge technical trade-offs create moderate friction against full automation.
Cost vs. human wageclaude-haiku-4-5-202510011/5The cost of AI assistance for managing vendor communication (drafting, summarization, documentation) is comparable to or exceeds the loaded hourly cost of a research scientist performing this task themselves, given the human judgment required to assess product fit.
Cost vs. human wageclaude-sonnet-52/5AI could reduce prep time (e.g., research summaries), but the core conferring and decision-making still requires a paid scientist, so total cost savings are modest relative to human wage.
Technical feasibility todayclaude-haiku-4-5-202510012/5While AI can draft emails or summarize vendor materials, no deployed product reliably conducts the full vendor evaluation conference independently. Vendor interactions require real-time adaptation, trust-building, and authority that AI systems today do not possess at production scale.
Technical feasibility todayclaude-sonnet-51/5No deployed product autonomously confers with vendors to negotiate equipment specifications or customize product lines for scientific users; this remains a human relationship-management activity.

Perform laboratory procedures following protocols including deoxyribonucleic acid (DNA) sequencing, cloning and extraction, ribonucleic acid (RNA) purification, or gel electrophoresis.

19

CI 7–30 · exposure 8 · augmentation 50 · importance 4.3/5 · click for rater detail

Panel ratings per dimension for this task
DimensionModelRatingRationale
Sector adoption velocityclaude-haiku-4-5-202510012/5Adoption of lab automation is slow and limited to well-funded research institutions and large pharma; most academic and clinical labs still rely on manual procedures due to cost, regulatory friction, and task complexity.
Sector adoption velocityclaude-sonnet-52/5Lab automation and robotics adoption in biology is gradual and capital-intensive, with AI itself playing a minor role in the physical execution of these procedures compared to digital-first sectors.
Augmentation potentialclaude-haiku-4-5-202510013/5AI assists moderately through protocol optimization, predictive quality checks, and data analysis of sequencing/gel results, but the physical execution of procedures remains human-dependent, limiting transformative augmentation.
Augmentation potentialclaude-sonnet-53/5AI can assist with protocol design, troubleshooting, data interpretation from sequencing or electrophoresis results, and documentation, but does not touch the hands-on execution of the procedures themselves.
Task automatabilityclaude-haiku-4-5-202510012/5While AI can assist in protocol design and data analysis, the core task requires hands-on laboratory work (sample handling, equipment operation, physical manipulation) that current systems cannot perform end-to-end. AI cannot reliably handle the dexterity, sensory feedback, and real-time adjustments needed for DNA/RNA extraction, gel loading, or equipment operation.
Task automatabilityclaude-sonnet-51/5This requires physical manipulation of biological samples, equipment operation, and fine motor skills that current AI systems cannot perform; lab automation robots exist but are not general AI and require significant separate infrastructure.
Adoption barriersclaude-haiku-4-5-202510014/5Strong barriers include regulatory requirements (especially for clinical/diagnostic labs), institutional biosafety protocols, quality assurance mandates for data integrity, and the need for trained human operators to ensure compliance and troubleshoot failures.
Adoption barriersclaude-sonnet-52/5No licensing requirement specifically bars automation of these procedures, but physical safety, sample integrity, and quality control create practical friction against non-human execution.
Cost vs. human wageclaude-haiku-4-5-202510012/5Specialized lab equipment and robotic systems (when available) are capital-intensive and require skilled technician oversight, making the all-in cost comparable to or higher than a trained lab technician's wage for most research-scale applications.
Cost vs. human wageclaude-sonnet-51/5AI cannot substitute for physical lab work at all, so there is no meaningful cost comparison for the task as stated; any automation would require costly robotic lab infrastructure, not standard AI inference costs.
Technical feasibility todayclaude-haiku-4-5-202510011/5No deployed autonomous systems reliably perform these wet-lab procedures in production settings. Robotic liquid handlers exist but require significant setup and human oversight; they are not off-the-shelf solutions that perform the full task independently.
Technical feasibility todayclaude-sonnet-51/5No AI product performs wet-lab procedures like pipetting, gel loading, or physical sample handling; automation here is via specialized robotics, not AI systems, and remains research/pilot-stage even then.

Coordinate molecular or cellular research activities with scientists specializing in other fields.

11

CI 5–16 · exposure 0 · augmentation 38 · importance 3.3/5 · click for rater detail

Panel ratings per dimension for this task
DimensionModelRatingRationale
Sector adoption velocityclaude-haiku-4-5-202510011/5Research institutions remain among the most conservative sectors for task automation, with strong cultural and organizational resistance to replacing human scientific judgment and interpersonal coordination.
Sector adoption velocityclaude-sonnet-52/5Academic and biotech research settings adopt AI tools slowly for high-level coordination tasks, though usage of AI for literature review or data analysis is growing.
Augmentation potentialclaude-haiku-4-5-202510012/5AI could modestly assist with scheduling, summarizing prior conversations, or drafting meeting agendas, but cannot materially augment the core judgment and relationship work that coordination demands.
Augmentation potentialclaude-sonnet-53/5AI can help summarize cross-disciplinary literature, draft communications, and organize shared data, meaningfully aiding coordination even though it doesn't replace the human role.
Task automatabilityclaude-haiku-4-5-202510011/5Coordinating cross-disciplinary research requires complex social navigation, judgment about scientific priorities, interpersonal negotiation, and real-time adaptive communication—none of which current AI systems can reliably perform end-to-end.
Task automatabilityclaude-sonnet-51/5Coordinating research across scientists requires relationship-building, scientific judgment, and negotiation of priorities that current AI cannot autonomously perform end-to-end.
Adoption barriersclaude-haiku-4-5-202510014/5Research collaboration is inherently human-centered; institutional cultures, funding agreements, and scientific accountability structures all expect and require human decision-makers and relationships. Liability and trust asymmetries are substantial.
Adoption barriersclaude-sonnet-53/5No licensing barrier per se, but organizational trust, accountability for research direction, and interpersonal dynamics create meaningful friction against full automation.
Cost vs. human wageclaude-haiku-4-5-202510011/5The cost and overhead of deploying AI for coordination (including human oversight, failed attempts, and required expertise to interpret outputs) far exceeds the cost of scientists directly communicating and coordinating with peers.
Cost vs. human wageclaude-sonnet-52/5AI can cheaply handle scheduling or note-taking, but the substantive scientific coordination still requires expensive human expert time, keeping overall cost comparable to or only slightly less than human-only coordination.
Technical feasibility todayclaude-haiku-4-5-202510011/5No deployed product reliably orchestrates multi-party scientific coordination, conflict resolution, or collaborative research planning in production environments. AI lacks the contextual understanding and relational continuity this role demands.
Technical feasibility todayclaude-sonnet-51/5No deployed product manages cross-disciplinary research coordination; existing tools only assist with scheduling or communication logistics, not substantive scientific coordination.

Provide scientific direction for project teams regarding the evaluation or handling of devices, drugs, or cells for in vitro and in vivo disease models.

6

CI 5–7 · exposure 0 · augmentation 63 · importance 4.0/5 · click for rater detail

Panel ratings per dimension for this task
DimensionModelRatingRationale
Sector adoption velocityclaude-haiku-4-5-202510011/5Academic and biotech sectors have adopted AI for supporting analysis and literature mining, but actual replacement or autonomous direction of project teams remains extremely rare. Institutional conservatism, regulatory caution, and the centrality of human expertise to funding and accountability slow adoption.
Sector adoption velocityclaude-sonnet-52/5Biotech/pharma R&D is adopting AI tools for specific subtasks (data analysis, literature review) but leadership and scientific direction functions show minimal displacement; adoption in wet-lab science remains slower than in pure information sectors.
Augmentation potentialclaude-haiku-4-5-202510013/5AI can meaningfully assist by retrieving relevant literature, summarizing disease models, flagging experimental design considerations, or analyzing omics data—augmenting the biologist's ability to synthesize information. However, the core directional judgment remains the human's responsibility.
Augmentation potentialclaude-sonnet-54/5AI can meaningfully assist by synthesizing literature, suggesting experimental designs, flagging risks, and analyzing prior data, substantially aiding the scientist's decision-making while they retain directive control.
Task automatabilityclaude-haiku-4-5-202510011/5This task requires deep scientific judgment, experimental design decisions, and contextual knowledge of disease biology that current AI cannot reliably perform end-to-end. AI can assist with literature review or hypothesis generation, but the directional leadership and accountability for model selection and validation remain fundamentally human responsibilities.
Task automatabilityclaude-sonnet-51/5This requires deep scientific judgment, novel hypothesis generation, and integration of tacit lab experience to direct research teams; current AI cannot autonomously provide reliable scientific leadership for experimental design in disease models.
Adoption barriersclaude-haiku-4-5-202510014/5Strong barriers exist: regulatory oversight (FDA, EMA, institutional review boards), institutional liability for flawed experimental direction, professional licensing and credentialing requirements for scientific leadership, and organizational norms requiring human expert accountability for research direction decisions.
Adoption barriersclaude-sonnet-54/5Providing scientific direction on drug/cell handling in disease models carries significant liability, regulatory oversight (IACUC, GLP, FDA-adjacent), and requires credentialed expertise, creating strong barriers to substitution.
Cost vs. human wageclaude-haiku-4-5-202510011/5The cost of deploying AI for this task, combined with required expert oversight and validation, exceeds the loaded wage of a skilled molecular biologist, especially given the high cost of errors in experimental design and model selection.
Cost vs. human wageclaude-sonnet-51/5There is no viable AI substitute performing this leadership/direction function, so cost comparison favors the human expert entirely; any AI cost would be additive to, not replacing, the biologist's role.
Technical feasibility todayclaude-haiku-4-5-202510011/5No deployed AI product reliably provides independent scientific direction for selecting and handling complex biological devices, drugs, or cells in disease models. While AI can suggest literature or analyze data, production systems do not autonomously direct project teams or validate experimental approaches.
Technical feasibility todayclaude-sonnet-51/5No deployed product exists that provides scientific direction to research teams for in vitro/in vivo model handling; this remains firmly in the research/consultative domain requiring expert human oversight.

Supervise technical personnel and postdoctoral research fellows.

4

CI 0–7 · exposure 0 · augmentation 25 · importance 4.0/5 · click for rater detail

Panel ratings per dimension for this task
DimensionModelRatingRationale
Sector adoption velocityclaude-haiku-4-5-202510011/5No measurable adoption of AI systems for primary supervisory roles exists in research institutions; supervision remains entirely human-driven across the sector.
Sector adoption velocityclaude-sonnet-52/5Academic and research lab settings adopt AI tools slowly for administrative tasks, and supervisory roles are not being displaced by AI systems currently.
Augmentation potentialclaude-haiku-4-5-202510012/5AI tools might assist with scheduling, data aggregation for performance metrics, or documentation, but they offer minimal augmentation to the core supervisory and mentoring functions themselves.
Augmentation potentialclaude-sonnet-52/5AI can help with scheduling, drafting performance reviews, or tracking project progress, but offers only marginal assistance to the core supervisory relationship.
Task automatabilityclaude-haiku-4-5-202510011/5Supervision of research personnel fundamentally requires real-time judgment, conflict resolution, mentoring, and adaptive communication—tasks requiring human empathy and contextual understanding that current AI systems cannot reliably replicate end-to-end.
Task automatabilityclaude-sonnet-51/5Supervising staff and postdocs requires interpersonal leadership, mentorship, performance evaluation, and contextual judgment that current AI cannot perform end-to-end.
Adoption barriersclaude-haiku-4-5-202510015/5Legal, organizational, and ethical requirements mandate that human supervisors directly oversee personnel management, performance reviews, and mentoring—supervisory authority cannot be delegated to AI systems.
Adoption barriersclaude-sonnet-54/5Personnel management involves legal responsibilities (HR compliance, evaluations, career mentorship) that require a human supervisor with authority and accountability.
Cost vs. human wageclaude-haiku-4-5-202510011/5AI systems cannot currently perform supervisory duties, so comparative cost analysis is not applicable; human supervisors remain mandatory.
Cost vs. human wageclaude-sonnet-51/5There is no viable AI substitute for supervisory labor, so cost comparison favors the human entirely.
Technical feasibility todayclaude-haiku-4-5-202510011/5No deployed product reliably manages personnel supervision, performance feedback, or mentoring relationships; these remain human-dependent functions in all real organizations.
Technical feasibility todayclaude-sonnet-51/5No deployed product manages personnel supervision autonomously; this remains a human management function in research labs.

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.