Lawyers

23-1011.00
Median wage $159,670/yr754,500 employed (US)Rank #661 of 923 scored · top 72% by substitution

Represent clients in criminal and civil litigation and other legal proceedings, draw up legal documents, or manage or advise clients on legal transactions. May specialize in a single area or may practice broadly in many areas of law.

Sub-scores

0–100 · band = confidence interval from rater disagreement

Substitution20
Exposure20
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%19

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

Technical feasibility todayw 20%22

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

Cost vs. human wagew 15%23

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

Adoption barriersw 20%inverted — strong barriers lower the score12

panel mean rating 4.5/5 (barrier strength) → substitution pressure 12/100

Sector adoption velocityw 10%33

panel mean rating 2.3/5 → substitution pressure 33/100

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.

Prepare, draft, and review legal documents, such as wills, deeds, patent applications, mortgages, leases, and contracts.

49

CI 49–49 · exposure 50 · augmentation 88 · importance 4.2/5 · click for rater detail

Panel ratings per dimension for this task
DimensionModelRatingRationale
Sector adoption velocityclaude-haiku-4-5-202510013/5Legal tech adoption is growing (contract review, document automation in corporate legal departments), but deployment remains concentrated in large firms and in-house teams. Solo and small practitioners lag significantly; full AI-driven document handling remains largely pilot-stage rather than mainstream production use.
Sector adoption velocityclaude-sonnet-53/5Legal services is a professional/information-sector field with growing AI adoption in drafting and review, but conservative firm culture, client trust concerns, and liability exposure keep production-scale deployment moderate rather than fast.
Augmentation potentialclaude-haiku-4-5-202510014/5AI substantially assists lawyers by drafting initial versions, flagging missing clauses, and suggesting revisions, allowing attorneys to focus on legal strategy and risk assessment. This augmentation is actively deployed and measurably improves productivity on document-heavy work while preserving the attorney's judgment role.
Augmentation potentialclaude-sonnet-55/5AI substantially speeds up first-draft generation, clause suggestion, and review/redlining for lawyers, who remain in the loop for judgment, negotiation, and final approval.
Task automatabilityclaude-haiku-4-5-202510013/5AI can draft routine legal documents (simple wills, leases, contracts) with significant templates and structure, achieving meaningful time savings on document generation. However, complex documents requiring jurisdiction-specific nuance, cross-referencing, and tailored risk analysis still require substantial human review and editing, limiting end-to-end automation below the 50% threshold for diverse practice contexts.
Task automatabilityclaude-sonnet-53/5AI can draft first-pass versions of standard documents like wills, leases, and contracts quickly, but complex or high-stakes documents (patents, negotiated contracts) still require substantial attorney review, editing, and judgment, limiting full end-to-end time savings to roughly half the task.'
Adoption barriersclaude-haiku-4-5-202510014/5Significant barriers exist: attorneys must sign off on legal documents for liability and malpractice protection, and many jurisdictions require a licensed attorney to verify compliance. Client preference for human counsel and professional norms around trust also slow substitution, though these are organizational rather than hard regulatory barriers.
Adoption barriersclaude-sonnet-54/5Practicing law and rendering legal advice requires bar licensure, and unauthorized practice of law statutes plus malpractice liability create strong barriers to full AI substitution for document preparation that carries legal effect.
Cost vs. human wageclaude-haiku-4-5-202510014/5AI document generation costs (via APIs or SaaS platforms) run $0.01–$0.10 per document after setup, while junior lawyers bill $150–$300/hour for similar work. The cost advantage strongly favors AI for routine document drafting and initial review, approaching an order of magnitude difference.
Cost vs. human wageclaude-sonnet-54/5AI drafting tools cost a small fraction of billable attorney hours for template-based documents, though oversight and review costs reduce the savings somewhat below a full order-of-magnitude for complex documents.
Technical feasibility todayclaude-haiku-4-5-202510013/5Products like LawGeex, Kira Systems, and general LLM-powered document generators perform narrow document generation and clause review in production, but with known error rates and scope limitations. They work reliably for standardized document types but falter on novel legal issues or multi-jurisdiction complexity, preventing a 5 rating.
Technical feasibility todayclaude-sonnet-53/5Products like legal drafting assistants and contract-review tools (e.g., Harvey, Spellbook, LawGeex) are deployed at some firms, but reliability varies by document complexity and jurisdiction, and human lawyers still finalize and sign off on most outputs.

Study Constitution, statutes, decisions, regulations, and ordinances of quasi-judicial bodies to determine ramifications for cases.

37

CI 28–47 · exposure 38 · augmentation 88 · importance 4.0/5 · click for rater detail

Panel ratings per dimension for this task
DimensionModelRatingRationale
Sector adoption velocityclaude-haiku-4-5-202510013/5Large law firms and corporate legal departments have begun piloting AI-assisted research tools, but adoption remains cautious and limited to supplementary roles. Most smaller practices and in-house counsel have not materially displaced human research work with automation.
Sector adoption velocityclaude-sonnet-54/5Legal services is a fast-adopting professional sector, with major firms and vendors rapidly integrating AI-based legal research and analysis tools into everyday practice.
Augmentation potentialclaude-haiku-4-5-202510014/5AI-powered legal research tools significantly augment lawyer productivity by rapidly surfacing relevant documents, statutes, and precedents, enabling lawyers to focus on higher-level analysis and strategic reasoning about ramifications for their cases.
Augmentation potentialclaude-sonnet-55/5AI dramatically speeds up locating and summarizing relevant statutes, case law, and regulatory provisions, significantly boosting lawyer productivity while the lawyer retains interpretive judgment.
Task automatabilityclaude-haiku-4-5-202510012/5While AI excels at retrieval and summarization of legal documents, determining ramifications for specific cases requires nuanced legal judgment, contextual understanding of precedent application, and strategic reasoning about case outcomes. Current systems cannot reliably perform this integrative analysis end-to-end at the quality level lawyers require.
Task automatabilityclaude-sonnet-53/5AI legal research tools can quickly surface and summarize relevant statutes, case law, and regulations, but synthesizing ramifications for a specific case still requires substantial attorney judgment and verification, limiting full end-to-end automation.
Adoption barriersclaude-haiku-4-5-202510014/5Bar admission and malpractice liability create substantial legal barriers: an attorney must ultimately verify and take responsibility for case strategy derived from this analysis. Client expectations and professional ethics rules require human expert judgment on legal ramifications.
Adoption barriersclaude-sonnet-54/5Legal analysis and case strategy determinations generally must be performed or certified by a licensed attorney, with malpractice liability creating strong incentives to keep humans in the loop.
Cost vs. human wageclaude-haiku-4-5-202510012/5Premium legal research platforms cost thousands annually per user, and lawyer oversight of AI outputs remains necessary, making the all-in cost comparable to or exceeding the hourly cost of junior associate research work on narrowly scoped issues.
Cost vs. human wageclaude-sonnet-53/5AI research subscriptions are cheaper per query than billable attorney hours, but the need for licensed oversight and verification of citations narrows the effective cost advantage.
Technical feasibility todayclaude-haiku-4-5-202510012/5AI legal research tools (e.g., LexisNexis+, Westlaw AI) exist and assist with document retrieval and summarization, but none reliably perform the full task of independently determining case ramifications without human review and judgment. Products show material limitations in reasoning about novel applications and client-specific contexts.
Technical feasibility todayclaude-sonnet-53/5Products like Westlaw Edge, Lexis+ AI, and Harvey are deployed in law firms for legal research, but they still exhibit hallucination risks and require careful lawyer review, so reliability is moderate rather than fully mature.

Search for and examine public and other legal records to write opinions or establish ownership.

35

CI 25–45 · exposure 42 · augmentation 88 · importance 3.7/5 · click for rater detail

Panel ratings per dimension for this task
DimensionModelRatingRationale
Sector adoption velocityclaude-haiku-4-5-202510012/5Adoption of AI for legal research assistance is growing in large law firms but remains cautious and experimental; small practices lag, and regulatory uncertainty about AI-generated opinions slows widespread deployment compared to other professional sectors.
Sector adoption velocityclaude-sonnet-53/5Legal services are adopting AI research tools at a moderate but accelerating pace, with pilots and partial production use common at large firms, while smaller practices and title/records work lag behind faster-digitizing sectors like finance.
Augmentation potentialclaude-haiku-4-5-202510014/5AI excels at accelerating legal research, identifying relevant cases, and summarizing documents, allowing attorneys to focus on analysis and opinion-writing; this assistive capability is measurably transforming lawyer productivity in the research phase while the attorney retains control.
Augmentation potentialclaude-sonnet-55/5AI dramatically speeds up searching, filtering, and summarizing large volumes of public records and case law, letting lawyers focus judgment and drafting on a much smaller curated set of relevant materials.
Task automatabilityclaude-haiku-4-5-202510012/5Legal research can be partially automated with AI tools (case law searches, document review), but synthesizing findings into legally sound opinions requires human judgment, ethical responsibility, and jurisdiction-specific interpretation that current systems cannot reliably perform end-to-end at equal quality.
Task automatabilityclaude-sonnet-53/5AI can rapidly search, summarize, and cross-reference legal and public records, and draft preliminary opinions, but establishing ownership often requires verifying original documents, resolving ambiguities, and applying judgment that current systems cannot fully replace end-to-end.
Adoption barriersclaude-haiku-4-5-202510015/5Lawyers must be licensed professionals, and opinions on ownership and legal matters require attorney sign-off and malpractice liability; regulatory and ethical rules mandate that attorneys—not AI alone—bear responsibility for the legal adequacy of work product.
Adoption barriersclaude-sonnet-54/5Legal opinions and formal ownership determinations often require an attorney's signature, malpractice liability attaches to errors, and many jurisdictions mandate licensed practitioners for such conclusions, creating strong professional and legal barriers to full automation.
Cost vs. human wageclaude-haiku-4-5-202510012/5While AI research tools reduce time on preliminary searches, the cost of AI integration, quality oversight, and attorney correction often approaches or exceeds the saved labor, especially given the liability cost of errors in legal opinions.
Cost vs. human wageclaude-sonnet-53/5AI tools substantially cut research time and cost compared to billed attorney hours, but licensing fees, integration, and required human review of ownership conclusions keep the ratio only moderately favorable rather than an order-of-magnitude cheaper for the full task.
Technical feasibility todayclaude-haiku-4-5-202510013/5Legal research products (LexisNexis with AI, ChatGPT plugins for legal searches) exist and function in production, but they still produce errors, miss relevant precedent, and cannot independently write opinions without significant attorney review and rewriting.
Technical feasibility todayclaude-sonnet-53/5Legal research and document review products (e.g., Westlaw AI, Casetext, Lexis+) are deployed in production and reliably assist with record search and summarization, but title/ownership determinations still commonly involve manual verification and human sign-off due to error sensitivity.

Analyze the probable outcomes of cases, using knowledge of legal precedents.

34

CI 31–36 · exposure 30 · augmentation 75 · importance 4.4/5 · click for rater detail

Panel ratings per dimension for this task
DimensionModelRatingRationale
Sector adoption velocityclaude-haiku-4-5-202510013/5Large law firms and corporations are piloting AI legal research and outcome prediction tools, but most are still in augmentation mode with attorney sign-off required. Adoption is faster in information-dense sectors (finance, IP) than in litigation with jury unpredictability.
Sector adoption velocityclaude-sonnet-53/5Legal services are a professional/information sector with growing AI tool adoption (research, drafting, analytics), but outcome prediction specifically remains in pilot/early-adoption stages rather than deep production use.
Augmentation potentialclaude-haiku-4-5-202510014/5AI tools that surface relevant precedents, identify patterns, and flag procedural risks substantially accelerate attorney preparation for outcome analysis. Lawyers retain judgment but move faster through precedent research and initial case assessment with AI assistance.
Augmentation potentialclaude-sonnet-54/5AI substantially speeds up precedent research, case law synthesis, and drafting of preliminary risk assessments, meaningfully boosting lawyer productivity while the lawyer retains final judgment.
Task automatabilityclaude-haiku-4-5-202510012/5AI can retrieve and summarize legal precedents and identify relevant case law patterns, but predicting case outcomes requires integrating jurisdiction-specific nuance, judge behavior, jury dynamics, and novel fact patterns that current systems handle inconsistently. This does not yet meet the 50% time-savings-at-equal-quality threshold end-to-end.
Task automatabilityclaude-sonnet-52/5AI can surface relevant precedents and generate probabilistic outcome estimates, but nuanced case-outcome prediction requires integrating facts, jurisdictional strategy, and judgment that current systems cannot reliably replicate end-to-end at equal quality.'
Adoption barriersclaude-haiku-4-5-202510014/5Lawyers must remain responsible for case strategy and outcome predictions to clients and courts; malpractice liability and professional conduct rules create high legal friction against full automation. Attorney work product privilege and ethical duties to clients further restrict autonomous deployment.
Adoption barriersclaude-sonnet-54/5Professional responsibility rules require licensed attorneys to exercise independent legal judgment and sign off on case strategy advice, creating strong liability and regulatory barriers to full delegation.
Cost vs. human wageclaude-haiku-4-5-202510013/5AI-assisted legal research costs substantially less than associate time spent on precedent review, but the final outcome-prediction analysis still requires attorney oversight and validation, keeping total cost comparable to hiring mid-level legal staff for preliminary analysis.
Cost vs. human wageclaude-sonnet-53/5AI-assisted research and analytics tools are cheaper than billed associate hours for initial research, but human lawyer review and judgment remain necessary, keeping blended costs closer to parity.
Technical feasibility todayclaude-haiku-4-5-202510013/5Legal research platforms (Westlaw, LexisNexis, specialized AI legal tools) can reliably retrieve precedents and draft outcome summaries, but outcome prediction remains in limited production use with material error rates. Most deployed systems assist rather than autonomously predict.
Technical feasibility todayclaude-sonnet-52/5Legal AI products (e.g., litigation analytics tools) exist and are used for precedent research and rough outcome forecasting, but they have material error rates and are not trusted as standalone case outcome predictors in practice.

Perform administrative and management functions related to the practice of law.

33

CI 25–41 · exposure 30 · augmentation 75 · importance 3.6/5 · click for rater detail

Panel ratings per dimension for this task
DimensionModelRatingRationale
Sector adoption velocityclaude-haiku-4-5-202510012/5While some law firms use practice management software, adoption of AI-driven management automation remains slow and uneven; many firms continue manual processes, and the profession's risk-averse culture and regulatory constraints limit rapid rollout in production.
Sector adoption velocityclaude-sonnet-53/5Legal services are adopting AI tools for administrative efficiency (billing, scheduling, document management) at a moderate pace, with pilots and partial deployments common but full-scale automation of management functions rare.
Augmentation potentialclaude-haiku-4-5-202510014/5AI can significantly assist lawyers with administrative workflows—automating scheduling, calendaring, billing reminders, document organization, and basic reporting—substantially raising attorney productivity while the lawyer retains full control and responsibility.
Augmentation potentialclaude-sonnet-54/5AI tools significantly assist with scheduling, billing, document drafting, and practice analytics, meaningfully boosting productivity while lawyers retain overall managerial control.
Task automatabilityclaude-haiku-4-5-202510012/5Current AI can automate routine administrative tasks (scheduling, document filing, basic record-keeping) but cannot reliably manage the practice as a whole—which involves judgment calls on client priorities, risk assessment, and strategic resource allocation that require human oversight.
Task automatabilityclaude-sonnet-52/5This is a broad catch-all task covering billing, staffing, client management, and office administration, most of which requires judgment, negotiation, and human coordination that current AI cannot fully replace end-to-end.
Adoption barriersclaude-haiku-4-5-202510014/5Strong barriers exist: lawyers must maintain legal responsibility and fiduciary duty to clients; malpractice liability falls on the attorney; regulatory bodies (bar associations) require human oversight of client account management and practice operations; client confidentiality rules limit what can be delegated or automated.
Adoption barriersclaude-sonnet-53/5No strict licensing requirement for administrative tasks themselves, but law firm governance, confidentiality obligations, and partner-level decision-making create organizational friction against full automation.
Cost vs. human wageclaude-haiku-4-5-202510013/5Practice management software and AI-assisted administrative tools cost roughly comparable to or slightly less than junior administrative staff when factoring in integration and oversight, placing them near parity rather than offering decisive cost advantage.
Cost vs. human wageclaude-sonnet-52/5While some subtasks like time tracking or document organization can be cheaply automated, the overall management function still requires substantial human oversight, keeping blended costs closer to human levels.
Technical feasibility todayclaude-haiku-4-5-202510013/5Document management systems and basic practice management software exist and perform predictably, but comprehensive AI solutions that handle the full scope of administrative and management functions remain limited in deployment and typically require significant human intervention.
Technical feasibility todayclaude-sonnet-52/5Practice management software and AI-assisted billing/scheduling tools exist, but no deployed product performs the full range of legal administrative and management functions autonomously and reliably.

Evaluate findings and develop strategies and arguments in preparation for presentation of cases.

32

CI 28–36 · exposure 30 · augmentation 75 · importance 4.3/5 · click for rater detail

Panel ratings per dimension for this task
DimensionModelRatingRationale
Sector adoption velocityclaude-haiku-4-5-202510013/5Large law firms and corporate legal departments are actively piloting AI research and brief-drafting tools, but production deployment of AI-driven strategy evaluation remains limited and cautious. Adoption is growing but constrained by liability concerns and the high stakes of individual cases.
Sector adoption velocityclaude-sonnet-53/5Legal services are adopting AI research and drafting tools rapidly, but adoption for actual strategic reasoning remains in early pilot stages within a traditionally cautious profession.
Augmentation potentialclaude-haiku-4-5-202510014/5AI strongly augments lawyer productivity on research, case law analysis, and evidence organization, allowing attorneys to focus on strategic judgment and argument development. Tools like legal research AI and contract analysis platforms measurably improve speed and comprehensiveness while attorneys retain decision-making authority.
Augmentation potentialclaude-sonnet-54/5AI substantially speeds up research, precedent analysis, and argument drafting, giving lawyers a strong productivity boost while they retain ultimate strategic control.
Task automatabilityclaude-haiku-4-5-202510012/5AI can assist in analyzing case law, organizing findings, and suggesting arguments through legal research tools, but cannot reliably evaluate complex legal strategy, weigh competing precedents against novel fact patterns, or develop novel arguments requiring experienced judgment. The task requires contextual legal reasoning and strategic discretion that current AI systems cannot perform end-to-end with 50% time savings at equal quality.
Task automatabilityclaude-sonnet-52/5AI can help synthesize facts and suggest arguments but developing case strategy requires judgment, client-specific risk assessment, and creative advocacy that current systems cannot reliably perform end-to-end.
Adoption barriersclaude-haiku-4-5-202510014/5Lawyers must exercise independent professional judgment and sign off on case strategy; courts and ethics rules require attorney accountability for arguments presented. Client relationships and fiduciary duties create strong friction against full automation, and malpractice liability remains with the licensed attorney regardless of AI assistance.
Adoption barriersclaude-sonnet-54/5Legal strategy and case argumentation typically require a licensed attorney's professional judgment and signature, with malpractice liability and bar rules limiting full delegation to AI.
Cost vs. human wageclaude-haiku-4-5-202510012/5Current AI legal tools reduce overhead on research and document review but require significant attorney oversight to validate findings and ensure strategy soundness. The loaded cost of AI plus necessary attorney review approximates the cost of direct attorney work on evaluation and strategy, without clear savings.
Cost vs. human wageclaude-sonnet-53/5AI drafting assistance is cheap per query, but the human lawyer time still required for strategic judgment and review keeps overall cost roughly comparable rather than dramatically cheaper.
Technical feasibility todayclaude-haiku-4-5-202510013/5Legal research and document analysis products (e.g., LexisNexis AI, Westlaw's AI-Assisted Research) exist and perform narrowly on cite-finding and precedent retrieval, but no deployed system reliably evaluates overall case strategy or develops comprehensive arguments at the quality expected in litigation. Material limitations remain in handling novel legal issues and nuanced fact patterns.
Technical feasibility todayclaude-sonnet-52/5Legal AI research tools (e.g., CoCounsel, Harvey) can assist with issue-spotting and argument drafting, but no deployed product independently develops full litigation strategy reliably at scale.

Negotiate contractual agreements.

28

CI 23–32 · exposure 30 · augmentation 75 · importance 3.9/5 · click for rater detail

Panel ratings per dimension for this task
DimensionModelRatingRationale
Sector adoption velocityclaude-haiku-4-5-202510013/5Law firms are piloting AI-assisted contract review and drafting at scale, but actual negotiation by AI remains experimental. Adoption is faster in document screening and initial drafting; negotiation automation is still in early exploration.
Sector adoption velocityclaude-sonnet-53/5Legal services are adopting AI tools quickly for research and drafting, but adoption for the negotiation function itself remains nascent and mostly assistive.
Augmentation potentialclaude-haiku-4-5-202510014/5AI significantly assists lawyers by rapidly analyzing counterproposals, flagging commercial risks, suggesting language alternatives, and tracking negotiation positions—enabling faster, more thorough preparation. The lawyer remains the negotiator, but with substantially amplified analytical capacity.
Augmentation potentialclaude-sonnet-54/5AI can meaningfully help lawyers prepare for negotiations by drafting terms, summarizing precedent, flagging risks, and simulating counterparty positions, improving efficiency while the lawyer leads the actual negotiation.
Task automatabilityclaude-haiku-4-5-202510012/5While AI can generate initial contract drafts, flag non-standard terms, and identify deviations from templates, the negotiation phase—which requires strategic judgment, understanding of counterparty intent, creative problem-solving, and real-time decision-making—remains beyond current AI capabilities. Humans must lead negotiations; AI provides analysis only.
Task automatabilityclaude-sonnet-52/5Negotiation requires real-time strategic judgment, reading counterparties, and authority to bind a client, which current AI cannot perform end-to-end; AI can only assist with drafting and analysis portions.'
Adoption barriersclaude-haiku-4-5-202510015/5Contract negotiation on behalf of clients involves fiduciary and legal authority requirements; only licensed attorneys can bind parties or represent clients in material negotiations. Regulatory and professional rules create hard barriers to full automation.
Adoption barriersclaude-sonnet-54/5Contract negotiation on behalf of a client typically requires an authorized, licensed attorney with fiduciary and ethical responsibilities, and liability for bad terms falls squarely on a human professional.
Cost vs. human wageclaude-haiku-4-5-202510012/5AI-powered contract review can reduce document analysis costs, but the labor saved is preliminary work. The core negotiation task still requires highly compensated attorneys, and AI tools add licensing and integration overhead without eliminating the primary cost driver.
Cost vs. human wageclaude-sonnet-52/5Because AI cannot substitute for the negotiation itself, the comparison is really about supporting tools, which add cost on top of the lawyer's time rather than replacing it wholesale.
Technical feasibility todayclaude-haiku-4-5-202510013/5Contract analysis and drafting tools are deployed in law firms, but these handle analysis and generation, not live negotiation. No production system reliably conducts end-to-end contract negotiation autonomously; existing products support lawyers rather than replace them.
Technical feasibility todayclaude-sonnet-52/5No deployed product autonomously negotiates contracts on a lawyer's behalf; existing tools focus on contract review, redlining suggestions, and clause drafting rather than live negotiation.

Examine legal data to determine advisability of defending or prosecuting lawsuit.

27

CI 23–31 · exposure 25 · augmentation 75 · importance 4.2/5 · click for rater detail

Panel ratings per dimension for this task
DimensionModelRatingRationale
Sector adoption velocityclaude-haiku-4-5-202510013/5Large law firms and corporate legal departments are adopting AI-assisted discovery and research tools at moderate pace, but the strategic decision to sue or defend remains a core attorney function with slow substitution in practice.
Sector adoption velocityclaude-sonnet-53/5Legal services show growing adoption of AI research and review tools, but full deployment for strategic litigation decisions remains in pilot/assistive stages rather than widespread production autonomy.
Augmentation potentialclaude-haiku-4-5-202510014/5AI significantly augments this task by rapidly synthesizing case law, identifying factual gaps, and modeling outcome probabilities, enabling lawyers to make faster, better-informed go/no-go decisions while retaining full accountability for the choice.
Augmentation potentialclaude-sonnet-54/5AI substantially accelerates legal research, precedent analysis, and risk-factor identification, materially boosting lawyer productivity even though final advisability judgments remain human-led.
Task automatabilityclaude-haiku-4-5-202510012/5While AI can assist in extracting and summarizing legal facts, the judgment call about whether to defend or prosecute requires weighing nuanced strategic, financial, and ethical considerations that remain fundamentally human decisions. Current AI cannot reliably replicate the contextual judgment, client counseling, and risk assessment that define this task.
Task automatabilityclaude-sonnet-52/5AI can summarize case law, surface relevant precedent, and flag risk factors, but the ultimate judgment of whether to litigate involves client-specific strategy, risk tolerance, and professional judgment that current systems cannot reliably execute end-to-end.olti
Adoption barriersclaude-haiku-4-5-202510015/5Attorneys must take direct professional and ethical responsibility for the advice to pursue or abandon litigation; bar associations and courts hold the lawyer—not an AI—accountable. Licensing and fiduciary duty create hard legal barriers to full automation.
Adoption barriersclaude-sonnet-54/5Only licensed attorneys can render this kind of legal advice and bear malpractice liability; ethical rules and unauthorized practice of law statutes create strong barriers to full automation of this judgment task.
Cost vs. human wageclaude-haiku-4-5-202510013/5AI-powered document review and legal research can reduce the cost of the information-gathering phase, but the decision-making portion still requires attorney time, so overall cost savings are partial and offset by oversight needs.
Cost vs. human wageclaude-sonnet-52/5AI tools reduce research time and cost, but the overall task still requires substantial attorney oversight, verification, and liability-bearing judgment, keeping all-in cost closer to comparable rather than order-of-magnitude cheaper.
Technical feasibility todayclaude-haiku-4-5-202510012/5AI tools can review documents and flag relevant legal precedents, but no deployed product reliably makes the go/no-go prosecution or defense decision autonomously. Legal research assistants exist, but the strategic recommendation itself remains a gap requiring human expertise.
Technical feasibility todayclaude-sonnet-52/5Legal research and document review products (e.g., CoCounsel, Harvey) exist and are used in practice, but reliable production-grade tools that autonomously assess litigation advisability rather than support research are narrow and still error-prone with hallucination risk.

Interpret laws, rulings and regulations for individuals and businesses.

25

CI 19–31 · exposure 25 · augmentation 75 · importance 4.5/5 · click for rater detail

Panel ratings per dimension for this task
DimensionModelRatingRationale
Sector adoption velocityclaude-haiku-4-5-202510013/5Law firms are adopting AI research tools and contract review platforms at moderate pace, with pilots widespread but full production automation of interpretation tasks rare; large corporate legal departments lead adoption, while smaller and solo practices lag significantly.
Sector adoption velocityclaude-sonnet-53/5Law firms are professional-services adopters piloting AI research/drafting tools broadly, but full-scale reliance on AI for legal interpretation remains cautious and supervised.
Augmentation potentialclaude-haiku-4-5-202510014/5AI excels at accelerating research, flagging relevant precedents, and drafting initial summaries of legal positions, materially raising lawyer productivity on interpretation tasks while the attorney retains decision-making authority and client responsibility.
Augmentation potentialclaude-sonnet-54/5AI substantially speeds up locating relevant statutes, case law, and drafting first-pass interpretations, letting lawyers focus on judgment and client-specific application.
Task automatabilityclaude-haiku-4-5-202510012/5While AI can retrieve and summarize legal texts and highlight relevant statute sections, interpreting laws for specific client situations requires contextual judgment, precedent synthesis, and strategic advice that current systems cannot reliably deliver end-to-end without substantial human oversight, falling well short of the 50% time-saving bar.
Task automatabilityclaude-sonnet-52/5AI can draft summaries and identify relevant statutes/rulings quickly, but authoritative interpretation requires judgment, contextual application, and accountability that current systems cannot reliably replicate end-to-end.'
Adoption barriersclaude-haiku-4-5-202510015/5Bar admission, malpractice liability, fiduciary duty, and jurisdictional rules explicitly require a licensed attorney to interpret law and provide legal advice; unauthorized practice of law is a crime in all U.S. states, creating a hard regulatory barrier to full automation.
Adoption barriersclaude-sonnet-54/5Unauthorized practice of law rules and malpractice liability mean only licensed attorneys can formally interpret law and provide binding advice to clients.
Cost vs. human wageclaude-haiku-4-5-202510011/5Current legal AI tools cost hundreds to thousands per matter in licensing and integration, while the attorney time they save (often 10–30%) does not offset the loaded cost of even junior lawyer time ($150–400/hour), and client risk management still demands lawyer review.
Cost vs. human wageclaude-sonnet-53/5AI-assisted research can cut associate hours substantially, but the need for attorney review to catch errors narrows the net cost advantage to roughly comparable rather than order-of-magnitude cheaper.
Technical feasibility todayclaude-haiku-4-5-202510012/5Legal research tools (LexisNexis AI, Westlaw's AI-Assisted Research) exist and assist with text retrieval and case matching, but deployed products do not perform full interpretive work reliably in production; they reduce research time but require expert lawyers to validate, contextualize, and render binding advice.
Technical feasibility todayclaude-sonnet-52/5Legal AI tools (e.g., CoCounsel, Harvey) assist with research and drafting interpretive memos, but hallucination risk and lack of jurisdictional nuance mean they are not yet reliable standalone interpreters in production.

Gather evidence to formulate defense or to initiate legal actions by such means as interviewing clients and witnesses to ascertain the facts of a case.

24

CI 20–28 · exposure 20 · augmentation 63 · importance 4.4/5 · click for rater detail

Panel ratings per dimension for this task
DimensionModelRatingRationale
Sector adoption velocityclaude-haiku-4-5-202510012/5Law firms are experimenting with AI document review and intake forms, but actual adoption of AI-driven evidence gathering in production remains limited; most firms retain human-led interviews as standard practice.
Sector adoption velocityclaude-sonnet-53/5Legal services are adopting AI for research, drafting, and transcription at moderate pace, but the interviewing/fact-gathering function itself sees little direct AI deployment yet.
Augmentation potentialclaude-haiku-4-5-202510013/5AI can assist by drafting interview outlines, summarizing witness statements, flagging inconsistencies in testimony, and organizing evidence, improving attorney efficiency without displacing human judgment in evidence evaluation and strategy.
Augmentation potentialclaude-sonnet-54/5AI tools can transcribe interviews, summarize testimony, flag inconsistencies, suggest follow-up questions, and organize evidence, meaningfully boosting lawyer productivity while the lawyer remains the one conducting interviews.
Task automatabilityclaude-haiku-4-5-202510012/5While AI can assist with initial document review and fact extraction, gathering evidence via interviews requires nuanced judgment, rapport-building, and real-time adaptation to witness testimony that current AI cannot reliably perform end-to-end without significant human oversight and intervention.
Task automatabilityclaude-sonnet-52/5Interviewing clients and witnesses requires building rapport, reading nonverbal cues, adapting questioning strategy in real time, and exercising legal judgment about relevance and privilege—capabilities current AI lacks for reliable end-to-end execution.", "AI can assist with prep and transcription but cannot conduct the interview itself at equal quality.
Adoption barriersclaude-haiku-4-5-202510014/5Client interviews and witness depositions involve attorney-client privilege, rules of evidence, and professional duty; legal ethics and liability exposure create strong barriers to full automation, as attorneys remain responsible for evidence integrity and case strategy.
Adoption barriersclaude-sonnet-54/5Attorney-client privilege, confidentiality obligations, and the need for a licensed attorney to strategically direct fact-gathering and assess admissibility create strong professional and legal barriers to full substitution.
Cost vs. human wageclaude-haiku-4-5-202510012/5AI systems for initial intake can reduce some labor, but the task still requires skilled attorneys or paralegals to conduct substantive interviews, meaning cost savings are modest compared to the full loaded human wage.
Cost vs. human wageclaude-sonnet-52/5Since AI cannot perform the core interviewing function, the human lawyer's time is still required, so cost savings are limited to peripheral documentation tasks rather than the interview itself.
Technical feasibility todayclaude-haiku-4-5-202510012/5No deployed product reliably conducts client or witness interviews autonomously; AI chatbots can collect basic information but lack the judgment and credibility required to build a defensible evidentiary record for legal proceedings.
Technical feasibility todayclaude-sonnet-51/5No deployed product autonomously conducts witness/client interviews to gather evidence; this remains firmly a human-performed task with AI only in supporting roles like transcription.

Probate wills and represent and advise executors and administrators of estates.

24

CI 20–28 · exposure 25 · augmentation 75 · importance 3.9/5 · click for rater detail

Panel ratings per dimension for this task
DimensionModelRatingRationale
Sector adoption velocityclaude-haiku-4-5-202510013/5Law firms are experimenting with AI-assisted document automation and estate management tools, but adoption remains in pilot and early-production phases. Most probate work is still handled by lawyers with marginal AI assistance, not full automation or agent-driven workflows.
Sector adoption velocityclaude-sonnet-52/5Legal services overall show cautious AI adoption due to malpractice risk and ethical rules, and probate/estate practice in particular remains a traditional, document-heavy, in-person area with slow uptake.
Augmentation potentialclaude-haiku-4-5-202510014/5AI can meaningfully assist by automating document drafting, organizing estate assets, flagging tax and statutory red flags, and summarizing applicable laws by jurisdiction—all of which a lawyer uses to advise executors more efficiently and thoroughly while maintaining professional judgment and accountability.
Augmentation potentialclaude-sonnet-54/5AI tools significantly speed up drafting of wills, petitions, inventories, and legal research summaries, letting attorneys handle probate matters more efficiently while retaining oversight and judgment.
Task automatabilityclaude-haiku-4-5-202510012/5While AI can help draft probate documents and summarize estate information, the task requires interpreting complex statutory requirements, jurisdictional variations, and unique family circumstances. A human lawyer must still review, advise, and make judgment calls on contested claims or tax implications, limiting time savings to under 50% for the full task end-to-end.
Task automatabilityclaude-sonnet-52/5Probate involves document drafting and procedural filing that AI can assist with, but the core task requires client counseling, fiduciary judgment, court appearances, and legal representation that cannot be fully automated today.
Adoption barriersclaude-haiku-4-5-202510014/5Probate is heavily regulated and jurisdiction-specific; only a licensed attorney can represent executors in court, file pleadings, and provide legal advice. Liability exposure is high if errors lead to improper distributions or tax penalties, and many courts require counsel-prepared filings.
Adoption barriersclaude-sonnet-55/5Only licensed attorneys can represent clients in probate court and provide legal advice to executors/administrators; unauthorized practice of law statutes create a hard legal barrier to full automation.
Cost vs. human wageclaude-haiku-4-5-202510012/5AI-assisted document generation reduces some clerical cost, but probate work is already relatively commoditized with modest hourly rates in many markets. The integrated cost of AI tools plus mandatory lawyer review and custom advice remains comparable to or higher than basic human paralegal handling of routine estates.
Cost vs. human wageclaude-sonnet-52/5AI can cheaply draft routine probate paperwork, but the overall task still requires substantial attorney time for advising, representation, and liability-bearing judgment, keeping costs comparable to human-only delivery.
Technical feasibility todayclaude-haiku-4-5-202510012/5AI tools exist for document drafting and basic estate analysis, but no deployed product reliably handles the full scope of probate work—especially contested matters, counsel to executors, and regulatory compliance across jurisdictions. Products remain narrow and require significant lawyer oversight to avoid errors.
Technical feasibility todayclaude-sonnet-52/5Legal AI tools exist for drafting probate documents and summarizing estate law, but no deployed product independently represents or advises executors in production; human attorneys remain essential for judgment and court interaction.

Prepare legal briefs and opinions, and file appeals in state and federal courts of appeal.

23

CI 20–26 · exposure 25 · augmentation 75 · importance 3.9/5 · click for rater detail

Panel ratings per dimension for this task
DimensionModelRatingRationale
Sector adoption velocityclaude-haiku-4-5-202510012/5Law firms have adopted AI research and drafting tools, but these remain assistive; there is no measurable displacement of brief-writing work. Adoption is confined to pilot projects and optional tools; production-scale automation of appellate filing is not occurring.
Sector adoption velocityclaude-sonnet-53/5Legal services is a professional/information sector with growing AI tool adoption, but conservative, liability-averse firm culture and court scrutiny of AI-generated filings has kept actual production deployment moderate rather than fast.
Augmentation potentialclaude-haiku-4-5-202510014/5AI significantly augments attorney productivity by automating legal research, generating draft language, and identifying case precedents, allowing attorneys to focus on legal strategy and argumentation. These tools are increasingly integrated into law practice, though the human attorney remains the essential decision-maker.
Augmentation potentialclaude-sonnet-54/5AI substantially speeds up legal research, drafting initial arguments, summarizing case law, and formatting citations, meaningfully boosting attorney productivity while the lawyer retains final judgment and signs off.
Task automatabilityclaude-haiku-4-5-202510012/5Current AI systems can draft sections of legal briefs and assist with legal research and writing, but cannot independently prepare a complete, arguable brief or opinion that meets jurisdictional requirements without significant human review and revision. The task requires legal judgment, jurisdiction-specific procedural compliance, and advocacy strategy that AI cannot fully execute autonomously.
Task automatabilityclaude-sonnet-52/5AI can draft substantial portions of briefs and research legal opinions, but the task requires original legal strategy, judgment calls on argument framing, and formal court filing with attorney certification that cannot be fully automated end-to-end today.4o
Adoption barriersclaude-haiku-4-5-202510015/5A licensed attorney must prepare and sign briefs and opinions filed in appellate courts; this is a hard regulatory requirement. Courts impose strict procedural rules, and the attorney bears malpractice and professional responsibility liability for the content, making autonomous AI substitution legally impermissible.
Adoption barriersclaude-sonnet-55/5Only licensed attorneys can file briefs and appeals in courts of appeal; unauthorized practice of law rules and mandatory attorney signature/certification requirements make this a hard legal barrier to substitution.
Cost vs. human wageclaude-haiku-4-5-202510012/5AI tools reduce the cost of research and initial drafting, but a qualified attorney must still review, revise, and sign every brief or opinion, meaning labor cost savings are partial at best. The all-in cost per filing remains dominated by attorney time, not AI inference.
Cost vs. human wageclaude-sonnet-53/5AI drafting tools reduce research and first-draft time significantly, but the need for extensive attorney review, editing, and liability oversight keeps all-in costs only moderately below traditional billable-hour costs.
Technical feasibility todayclaude-haiku-4-5-202510012/5AI legal research and writing tools exist (e.g., document drafting assistants, case-law search engines), but no deployed product reliably produces court-filing-ready briefs or opinions without material human oversight. Products are confined to research assistance and drafting support; appellate filings must be reviewed and certified by licensed attorneys.
Technical feasibility todayclaude-sonnet-52/5Legal AI tools (e.g., Casetext, Harvey, CoCounsel) demonstrably assist with drafting and research in production at some firms, but hallucination risks and court sanctions for AI-generated errors show these tools are not yet reliable enough to perform the full task unsupervised.

Help develop federal and state programs, draft and interpret laws and legislation, and establish enforcement procedures.

23

CI 20–25 · 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 legislative bodies and law firms use AI for legal research and document drafting, actual deployment of AI to develop programs and establish enforcement procedures is minimal. Adoption remains in the pilot and research phase; there is no evidence of deep, production-scale substitution.
Sector adoption velocityclaude-sonnet-52/5Government and legislative drafting offices are slow-moving, bureaucratic, and cautious about AI use in binding legal text, resulting in limited production adoption despite general legal-tech growth.
Augmentation potentialclaude-haiku-4-5-202510013/5AI can assist lawyers and legislative drafters by rapidly generating precedent summaries, highlighting textual inconsistencies, and producing initial language drafts, materially raising productivity. However, augmentation is limited to parts of the task; final policy synthesis and enforcement design remain human-driven.
Augmentation potentialclaude-sonnet-54/5AI tools substantially assist with legal research, drafting language variations, comparing statutory text, and summarizing precedent, meaningfully speeding up the human-led process.
Task automatabilityclaude-haiku-4-5-202510012/5While AI can draft legislative language and identify legal precedents at scale, developing coherent federal/state programs and establishing enforcement procedures requires balancing competing policy goals, constitutional constraints, and political feasibility that demand human judgment. Current AI cannot reliably end-to-end substitute for this task with ≥50% time savings at equal quality.
Task automatabilityclaude-sonnet-52/5This task requires deep political judgment, stakeholder negotiation, and strategic drafting decisions that current AI cannot autonomously perform; AI can draft language but cannot independently develop policy or negotiate enforcement frameworks end-to-end.rationale
Adoption barriersclaude-haiku-4-5-202510015/5Developing federal and state legislation and enforcement procedures is constitutionally and statutorily reserved to elected officials and licensed attorneys. No AI system can legally sign off on or replace human legislative authority; this is a hard legal and democratic barrier.
Adoption barriersclaude-sonnet-54/5Drafting binding legislation and enforcement procedures typically requires licensed attorneys or government officials with legal authority to finalize and certify such work, creating strong professional and institutional barriers.
Cost vs. human wageclaude-haiku-4-5-202510012/5AI research and drafting assistance is cheap per token, but the integration overhead, human lawyer oversight, and need for senior judgment on policy substance means total cost per substantive legislative output remains comparable to or exceeds human lawyer labor.
Cost vs. human wageclaude-sonnet-52/5AI can cheaply generate draft text but the overall task involves substantial expert review, negotiation, and legal judgment, so all-in cost savings versus a lawyer's fully loaded output are modest.
Technical feasibility todayclaude-haiku-4-5-202510012/5AI legal research and drafting tools exist and assist with document generation, but no deployed product autonomously develops legislative programs or establishes enforcement procedures reliably. These tasks remain research-stage or narrow proof-of-concept; production systems do not handle the full scope.
Technical feasibility todayclaude-sonnet-52/5Legal drafting assistants and legislative research tools exist and are used to speed drafting, but no deployed product reliably develops full programs or enforcement schemes without extensive human legal expertise driving the process.

Advise clients concerning business transactions, claim liability, advisability of prosecuting or defending lawsuits, or legal rights and obligations.

19

CI 16–23 · exposure 25 · augmentation 75 · importance 4.3/5 · click for rater detail

Panel ratings per dimension for this task
DimensionModelRatingRationale
Sector adoption velocityclaude-haiku-4-5-202510012/5Legal services adoption of AI remains cautious and piecemeal. Firms use AI for document review and research acceleration, but autonomous client advisory remains uncommon; risk-averse organizational culture, regulatory uncertainty, and liability concerns slow deep automation in practice.
Sector adoption velocityclaude-sonnet-53/5Legal services are a professional-services sector with growing AI pilot adoption (research, drafting, contract review) but actual client advisory work remains conservatively adopted due to liability concerns.
Augmentation potentialclaude-haiku-4-5-202510014/5AI substantially augments lawyer productivity for research, contract drafting, due diligence, and precedent analysis. A lawyer can leverage AI-generated case summaries, risk assessments, and option analyses to advise clients faster and more thoroughly, with the human lawyer retaining final judgment and accountability.
Augmentation potentialclaude-sonnet-54/5AI substantially augments lawyers by summarizing case law, drafting risk analyses, and surfacing precedent, letting attorneys form and deliver advice faster while retaining ultimate judgment and liability.
Task automatabilityclaude-haiku-4-5-202510012/5While AI can assist with legal research, document review, and initial case analysis, providing comprehensive business advice requires judgment about client circumstances, risk tolerance, and strategy that depends on nuanced human interaction. Current AI cannot reliably handle the full advisory loop—understanding client context, weighing competing legal and business interests, and committing to a position—at quality parity with a lawyer.
Task automatabilityclaude-sonnet-52/5Legal advice requires synthesizing client-specific facts, judgment about risk tolerance, and accountability for consequences; AI can draft analysis but cannot reliably replace the full advisory judgment and client relationship end-to-end.
Adoption barriersclaude-haiku-4-5-202510015/5Strong legal and regulatory barriers exist: lawyers are licensed professionals, malpractice liability and fiduciary duty attach personally to the attorney, and bar ethics rules require that client advice come from a licensed lawyer who can be held accountable. Courts and regulators have not authorized AI to independently advise clients on legal matters.
Adoption barriersclaude-sonnet-55/5Giving legal advice is subject to unauthorized-practice-of-law statutes and bar licensing requirements; only licensed attorneys may formally advise clients, creating a hard legal barrier to automation.
Cost vs. human wageclaude-haiku-4-5-202510011/5Even accounting for AI's low marginal inference cost, the human oversight, liability exposure, and need for a qualified attorney to validate and take responsibility for advice means the all-in cost remains dominated by lawyer time. Substitution does not occur at meaningful cost advantage today.
Cost vs. human wageclaude-sonnet-52/5AI tools reduce research and drafting time but licensed attorney review, malpractice exposure, and client-facing counsel still require billed attorney time, keeping costs comparable rather than order-of-magnitude lower.
Technical feasibility todayclaude-haiku-4-5-202510012/5AI legal products (e.g., contract analysis tools, due diligence assistants) exist and are used in some firms, but they lack sufficient reliability and scope for independent client advisory work. Deployed systems still require lawyer review and cannot make final recommendations without human expertise, particularly for novel or complex scenarios involving liability or litigation strategy.
Technical feasibility todayclaude-sonnet-52/5Legal AI tools (Harvey, CoCounsel, etc.) assist with research and drafting memos but no deployed product independently advises clients on legal rights/strategy without attorney review and sign-off.

Confer with colleagues with specialties in appropriate areas of legal issue to establish and verify bases for legal proceedings.

16

CI 7–25 · exposure 13 · augmentation 50 · importance 3.8/5 · click for rater detail

Panel ratings per dimension for this task
DimensionModelRatingRationale
Sector adoption velocityclaude-haiku-4-5-202510012/5Legal services remain heavily human-gatekept despite digitization. Adoption of AI for independent collegial consultation and legal strategy decisions is minimal; most law firms use AI as a research augment rather than as an autonomous decision-maker on case foundations.
Sector adoption velocityclaude-sonnet-52/5While legal services are adopting AI for research and drafting, the specific interpersonal act of conferring with specialist colleagues is not a target of AI deployment and shows minimal adoption.
Augmentation potentialclaude-haiku-4-5-202510013/5AI can usefully assist lawyers in gathering precedents, identifying relevant specialists, and summarizing case law to inform collegial discussions. However, the core task—conferring and jointly establishing the legal basis—remains driven by human attorney judgment, limiting the transformative impact.
Augmentation potentialclaude-sonnet-53/5AI can assist by summarizing legal issues, surfacing relevant precedents, or preparing briefing materials before or after such consultations, improving efficiency without replacing the human exchange.
Task automatabilityclaude-haiku-4-5-202510012/5While AI can assist in legal research and identifying relevant precedents, the task requires nuanced collegial judgment, interpretation of complex legal strategy, and establishment of case foundations—activities that demand experienced legal expertise and human discretion. AI cannot reliably execute the full collaborative decision-making process independently.
Task automatabilityclaude-sonnet-51/5This is an interpersonal, collaborative professional consultation requiring judgment-sharing among licensed experts; AI cannot substitute for the act of conferring with colleagues to establish legal strategy.'
Adoption barriersclaude-haiku-4-5-202510014/5Strong barriers exist: legal liability for inadequate case foundations, professional responsibility rules requiring attorney sign-off on strategy decisions, and regulatory/malpractice asymmetries that make clients and firms reluctant to delegate case-foundation decisions to AI without human attorney verification.
Adoption barriersclaude-sonnet-54/5Legal practice involves licensing, professional responsibility, and liability concerns that require attorneys themselves to verify legal bases and consult peers, creating strong professional/regulatory barriers to automation.
Cost vs. human wageclaude-haiku-4-5-202510012/5AI legal research tools reduce research time but do not replace the senior attorney time required for collegial consultation, expert judgment, and verification. The loaded cost of human expertise combined with AI oversight remains comparable to or higher than using humans alone for this task.
Cost vs. human wageclaude-sonnet-51/5There is no AI product performing this collaborative task, so no cost comparison favors AI; the human interaction itself is the deliverable.
Technical feasibility todayclaude-haiku-4-5-202510012/5No deployed product reliably performs end-to-end collegial legal consultation and case-foundation verification. AI legal research tools exist but operate as assistants; they do not independently confer with specialists or establish verified bases for proceedings at production scale.
Technical feasibility todayclaude-sonnet-51/5No deployed product performs collegial legal consultation and joint verification of legal bases; this remains a human-to-human professional interaction.

Present and summarize cases to judges and juries.

7

CI 0–15 · exposure 8 · augmentation 63 · importance 4.4/5 · click for rater detail

Panel ratings per dimension for this task
DimensionModelRatingRationale
Sector adoption velocityclaude-haiku-4-5-202510011/5Legal presentation to courts is inherently human-facing and heavily regulated; adoption of autonomous AI is effectively zero. Legal tech adoption focuses on document review and research, not courtroom substitution, and regulatory and liability structures actively prevent such substitution.
Sector adoption velocityclaude-sonnet-52/5While legal services broadly are adopting AI for research and drafting, courtroom presentation itself sees virtually no AI deployment due to procedural and ethical constraints.
Augmentation potentialclaude-haiku-4-5-202510013/5AI can usefully assist lawyers by drafting case summaries, organizing evidence, suggesting argument structures, and preparing visual aids before trial. However, the human attorney remains fully in control of live presentation, making this assistive rather than transformative.
Augmentation potentialclaude-sonnet-54/5AI can help lawyers prepare arguments, summarize case law, draft outlines, and rehearse responses, meaningfully boosting preparation productivity even though the human still delivers the presentation.
Task automatabilityclaude-haiku-4-5-202510012/5While AI can draft case summaries and organize legal arguments, presenting to judges and juries requires real-time adaptation, persuasion, credibility-building, and courtroom presence that current systems cannot replicate. AI cannot reliably handle judicial questions, objections, or jury psychology in live adversarial settings at equal quality.
Task automatabilityclaude-sonnet-51/5Oral advocacy before judges and juries requires live persuasion, real-time adaptation, courtroom presence, and credibility that current AI cannot replicate end-to-end; no product performs this task itself.
Adoption barriersclaude-haiku-4-5-202510015/5Courtroom presentation is protected by strict licensing requirements (bar admission), ethical rules of conduct (Model Rules 3.3, 3.4), and constitutional rights to counsel. Only licensed attorneys can represent clients in court; no automation framework can substitute for human attorney presence and accountability.
Adoption barriersclaude-sonnet-55/5Only licensed attorneys (or pro se parties) may represent clients and speak in court; unauthorized practice of law rules make this a hard legal barrier against AI substitution.
Cost vs. human wageclaude-haiku-4-5-202510011/5The cost of AI infrastructure, legal knowledge integration, courtroom compliance systems, and mandatory human attorney oversight (both ethically and legally) would exceed the loaded cost of having a lawyer present directly. Liability and error costs for AI-led presentations are also prohibitively high.
Cost vs. human wageclaude-sonnet-52/5Even if AI drafted argument content, the delivery still requires a licensed attorney present in court, so there's minimal cost savings on the actual task of presenting to judges/juries.
Technical feasibility todayclaude-haiku-4-5-202510011/5No deployed product can end-to-end present cases to courts; this remains research-stage. AI tools exist for brief drafting and argument organization, but courtroom presentation—with its human judgment, ethical obligations, and immediate adaptation requirements—is not yet performed by deployed systems in production.
Technical feasibility todayclaude-sonnet-51/5No deployed product argues cases or delivers courtroom presentations in real trials; this remains firmly research/demo territory at best (e.g., experimental AI 'lawyers' have been blocked or withdrawn).

Negotiate settlements of civil disputes.

5

CI 3–7 · exposure 5 · augmentation 50 · importance 4.0/5 · click for rater detail

Panel ratings per dimension for this task
DimensionModelRatingRationale
Sector adoption velocityclaude-haiku-4-5-202510012/5Legal sectors are early in AI adoption and primarily use it for document review and research; negotiation itself remains almost entirely human-performed, with little evidence of AI agents displacing this function.
Sector adoption velocityclaude-sonnet-52/5Legal services adopt AI for research and drafting but negotiation itself remains almost entirely human-led, with slow structural change in this specific activity.
Augmentation potentialclaude-haiku-4-5-202510013/5AI can usefully assist lawyers by preparing settlement analyses, modeling outcomes, drafting terms, and organizing case data, allowing human negotiators to focus on strategy and persuasion, though the interaction remains lawyer-centric.
Augmentation potentialclaude-sonnet-53/5AI can help lawyers prepare by analyzing case value, precedent, and drafting settlement terms, usefully supporting but not replacing the negotiation itself.
Task automatabilityclaude-haiku-4-5-202510011/5Settlement negotiation requires real-time judgment about opposing parties' intentions, creative deal structuring, and strategic concessions tied to case risk—tasks demanding human judgment, trust-building, and accountability that current AI cannot perform end-to-end.
Task automatabilityclaude-sonnet-51/5Negotiation requires real-time strategic judgment, client authority, relationship dynamics, and adaptive persuasion that current AI cannot autonomously execute end-to-end.
Adoption barriersclaude-haiku-4-5-202510015/5Clients must be represented by a licensed attorney during settlement negotiations; this is a hard legal and fiduciary requirement in most jurisdictions, and unauthorized practice of law prohibits AI systems from negotiating on behalf of parties.
Adoption barriersclaude-sonnet-55/5Only licensed attorneys can bind clients and negotiate legal settlements; bar rules, fiduciary duty, and liability make this a hard-barrier task requiring authorized human representation.
Cost vs. human wageclaude-haiku-4-5-202510011/5AI tools (research, document generation) reduce some preparatory work, but the skilled negotiation itself—which is the core of this task—still requires a lawyer's time and cannot be replaced by cheaper AI inference.
Cost vs. human wageclaude-sonnet-51/5Since AI cannot perform the negotiation itself, there is no viable AI-only cost comparison; a human attorney remains necessary for the core task.
Technical feasibility todayclaude-haiku-4-5-202510012/5AI can assist with legal research and document drafting for settlements, but no deployed system can autonomously negotiate on behalf of clients; negotiation remains a human-to-human function where AI has only peripheral roles.
Technical feasibility todayclaude-sonnet-51/5No deployed product conducts actual settlement negotiations on behalf of a lawyer; AI is used at most for drafting or analysis support, not the negotiation act itself.

Select jurors, argue motions, meet with judges, and question witnesses during the course of a trial.

4

CI 0–7 · exposure 5 · augmentation 50 · importance 4.3/5 · click for rater detail

Panel ratings per dimension for this task
DimensionModelRatingRationale
Sector adoption velocityclaude-haiku-4-5-202510011/5Trial conduct automation is negligible in deployment. Legal tech adoption focuses on document review and research support, not courtroom replacement; the high stakes and regulatory barriers mean nearly zero displacement of trial lawyers.
Sector adoption velocityclaude-sonnet-52/5Legal services broadly are adopting AI for research and drafting, but courtroom advocacy itself sees essentially no AI deployment due to procedural and ethical constraints.
Augmentation potentialclaude-haiku-4-5-202510013/5AI can assist with case research, motion drafting suggestions, and witness preparation materials, raising lawyer productivity on preparation—but the core trial task (live argumentation, cross-examination, juror management) remains human-driven.
Augmentation potentialclaude-sonnet-53/5AI can help attorneys prepare questions, anticipate objections, and analyze juror data beforehand, but offers no real-time assistance during actual trial proceedings.
Task automatabilityclaude-haiku-4-5-202510011/5This task requires real-time courtroom argumentation, strategic judgment about juror suitability, and adaptive questioning of witnesses—all demand human judgment, persuasion, and legal accountability that current AI cannot reliably perform end-to-end. No AI system can autonomously conduct trials or cross-examinations at quality parity with licensed attorneys.
Task automatabilityclaude-sonnet-51/5Live courtroom advocacy—voir dire, motion argument, real-time witness examination—requires in-person judgment, adaptive rhetoric, and legal authorization no current AI can execute end-to-end.
Adoption barriersclaude-haiku-4-5-202510015/5This task is legally protected: only licensed attorneys can practice before courts, select jurors, and conduct trial proceedings. Judges and bar authorities enforce these requirements; automation is blocked by licensing law and professional regulation.
Adoption barriersclaude-sonnet-55/5Only licensed attorneys may argue motions and examine witnesses in court; this is a hard legal/regulatory barrier with direct liability and bar admission requirements.
Cost vs. human wageclaude-haiku-4-5-202510011/5Given the task's complexity and liability exposure, human lawyers remain far cheaper than any AI system that would need extensive oversight and integration to approach trial capability; current tools offer only marginal assistance on isolated components.
Cost vs. human wageclaude-sonnet-51/5Since AI cannot perform the task, there is no viable cost comparison; any attempt would require full human oversight negating savings.
Technical feasibility todayclaude-haiku-4-5-202510012/5No deployed product can autonomously conduct trial proceedings, select jurors, or argue motions before judges. Research prototypes exist for limited tasks like brief drafting aids, but nothing performs trial conduct reliably in production settings.
Technical feasibility todayclaude-sonnet-51/5No deployed product performs live trial advocacy; AI is used only for pre-trial prep like research or mock questioning, not the courtroom task itself.

Supervise legal assistants.

4

CI 0–7 · exposure 0 · augmentation 38 · importance 4.0/5 · click for rater detail

Panel ratings per dimension for this task
DimensionModelRatingRationale
Sector adoption velocityclaude-haiku-4-5-202510011/5No meaningful AI adoption for staff supervision exists because the task is fundamentally interpersonal and management-critical. Lawyers and firms have shown no movement toward AI-driven supervision, and such a shift would face severe organizational and legal resistance.
Sector adoption velocityclaude-sonnet-52/5Legal services broadly show slow-to-moderate AI adoption, and management/supervisory functions specifically remain almost untouched by automation tools.
Augmentation potentialclaude-haiku-4-5-202510012/5AI could marginally assist by summarizing assistant work products or flagging performance metrics, but these are peripheral to the core supervisory relationship of feedback, coaching, and accountability. The human supervisor remains essential and central.
Augmentation potentialclaude-sonnet-53/5AI can help lawyers track assistant workloads, review draft work product, and organize case management, offering moderate productivity support without replacing the supervisory judgment involved.
Task automatabilityclaude-haiku-4-5-202510011/5Supervising legal assistants requires nuanced human judgment about performance, development, motivation, and interpersonal dynamics. Current AI systems cannot autonomously manage, evaluate, or direct human staff, nor can they handle the context-dependent feedback and mentoring this task demands.
Task automatabilityclaude-sonnet-51/5Supervising people requires judgment, mentorship, delegation, and accountability that current AI systems cannot perform end-to-end; no viable path to autonomous supervision of staff exists today.
Adoption barriersclaude-haiku-4-5-202510015/5Legal supervision of staff is a core management responsibility with implicit organizational accountability, employment law obligations, and liability for decisions affecting employees. Only a human supervisor can legally and ethically perform this function.
Adoption barriersclaude-sonnet-54/5Supervisory responsibility carries professional liability and ethical obligations (e.g., bar rules requiring attorney oversight of non-lawyer staff) that legally require a human lawyer in this role.
Cost vs. human wageclaude-haiku-4-5-202510011/5AI cannot perform this task at all, making cost comparison moot. A human supervisor must be in place, and AI offers no substitution pathway, so the ratio is undefined in automation's favor.
Cost vs. human wageclaude-sonnet-51/5There is no AI substitute performing this managerial task, so cost comparison favors the human lawyer by default.
Technical feasibility todayclaude-haiku-4-5-202510011/5No deployed product can reliably supervise, evaluate, or manage human employees. This remains entirely within human management domains; AI has no production systems that autonomously oversee staff performance or replace supervisory functions.
Technical feasibility todayclaude-sonnet-51/5No deployed product manages or supervises human legal assistants; AI tools at best assist with task tracking or document review, not personnel supervision.

Represent clients in court or before government agencies.

1

CI 0–3 · exposure 0 · augmentation 38 · importance 4.3/5 · click for rater detail

Panel ratings per dimension for this task
DimensionModelRatingRationale
Sector adoption velocityclaude-haiku-4-5-202510011/5Courtroom representation has shown minimal AI adoption. Lawyers remain gatekeepers, bar associations regulate practice strictly, and liability concerns prevent substitution. Adoption remains at the research or pilot stage with no meaningful production displacement.
Sector adoption velocityclaude-sonnet-52/5Legal services broadly are seeing growing AI pilot use for research and drafting, but the courtroom advocacy function itself remains essentially untouched by adoption trends.
Augmentation potentialclaude-haiku-4-5-202510012/5AI assists attorneys with case research and brief drafting before court, but offers limited augmentation during live advocacy itself. Real-time assistance during oral argument or cross-examination remains narrow; the core task of representing clients in the courtroom remains fundamentally human-dependent.
Augmentation potentialclaude-sonnet-53/5AI can help lawyers prepare arguments, research precedent, draft briefs, and rehearse oral arguments, meaningfully boosting preparation productivity even though it cannot appear in their place.
Task automatabilityclaude-haiku-4-5-202510011/5Court representation requires live oral advocacy, real-time legal judgment calls, cross-examination, and legal accountability that current AI cannot perform. While AI can draft documents or suggest arguments, it cannot substitute for the lawyer's presence, judgment, and licensed advocacy role in contested proceedings.
Task automatabilityclaude-sonnet-51/5Courtroom representation requires live oral argument, real-time adaptation, credibility judgments, and physical presence before a judge or agency official, none of which current AI can perform end-to-end.
Adoption barriersclaude-haiku-4-5-202510015/5Courtroom representation is gated by bar licensing and law requiring a licensed human attorney to appear, sign pleadings, and take legal responsibility. Courts and agencies have explicit rules mandating human attorney presence and conduct; no automation can bypass these hard legal requirements.
Adoption barriersclaude-sonnet-55/5Court appearances legally require a licensed attorney (or authorized pro se party); bar admission, rules of professional conduct, and court rules make this one of the most strongly gated tasks in any occupation.
Cost vs. human wageclaude-haiku-4-5-202510011/5Even if partial automation were feasible, a licensed attorney's billable hourly rate far exceeds the cost of AI inference and oversight combined. The task remains labor-intensive, and current AI cannot replace the bulk of attorney time spent on live representation.
Cost vs. human wageclaude-sonnet-51/5Since AI cannot perform the actual representation function, there is no viable AI-only cost basis to compare; any attempt would still require a licensed human attorney of record, so no cost savings accrue.
Technical feasibility todayclaude-haiku-4-5-202510011/5No deployed product performs courtroom representation or agency advocacy end-to-end. AI systems exist for legal research and document drafting, but no production system reliably argues cases, responds to judicial questions, or manages adversarial proceedings independently.
Technical feasibility todayclaude-sonnet-51/5No deployed product appears in courtrooms or before agencies as the representing advocate; AI is at most a research curiosity in this exact task, not a fielded product replacing counsel.

Present evidence to defend clients or prosecute defendants in criminal or civil litigation.

1

CI 0–3 · exposure 0 · augmentation 38 · importance 4.3/5 · click for rater detail

Panel ratings per dimension for this task
DimensionModelRatingRationale
Sector adoption velocityclaude-haiku-4-5-202510011/5Litigation practice remains among the most conservative sectors for AI adoption. Courtroom work is deeply relationship-driven, high-stakes, and subject to ethical rules and judicial gatekeeping that slow any technological substitution. Observed adoption is minimal.
Sector adoption velocityclaude-sonnet-52/5While legal services broadly are adopting AI for research and drafting, courtroom advocacy itself sees essentially no AI adoption due to procedural and licensing constraints.
Augmentation potentialclaude-haiku-4-5-202510012/5AI provides modest assistance in preparing evidence presentation (organizing exhibits, summarizing documents, suggesting rhetorical structures) but cannot augment the live courtroom performance itself. The human lawyer remains entirely in the loop and irreplaceable during the actual presentation.
Augmentation potentialclaude-sonnet-53/5AI can help lawyers prepare by organizing evidence, drafting arguments, and predicting case outcomes, but it does not directly enhance the live presentation itself.
Task automatabilityclaude-haiku-4-5-202510011/5Presenting evidence in court requires real-time judgment about admissibility, strategic disclosure, opponent response, and courtroom dynamics that demand human agency and accountability. Current AI cannot navigate the adversarial legal environment, judge credibility live, or make split-second decisions on case strategy that courts require.
Task automatabilityclaude-sonnet-51/5Live courtroom presentation of evidence requires real-time human judgment, adaptive argumentation, credibility assessment, and physical presence before a judge/jury that no current AI system can perform end-to-end.
Adoption barriersclaude-haiku-4-5-202510015/5Courtroom presentation of evidence is a core licensed practice of law; only admitted attorneys may conduct litigation and present evidence in court. Jurisdictional bar rules and courtroom procedure rules explicitly require a human attorney as representative, creating hard legal barriers.
Adoption barriersclaude-sonnet-55/5Only licensed attorneys admitted to the bar may represent clients and present evidence in court, a strict legal and ethical requirement that fully blocks AI substitution.
Cost vs. human wageclaude-haiku-4-5-202510011/5The cost of developing and deploying an AI system capable of courtroom litigation, combined with required human oversight and liability insurance, would far exceed the hourly rate of even senior trial counsel. Integration risk and error costs make this economically infeasible.
Cost vs. human wageclaude-sonnet-51/5Since AI cannot perform the core courtroom task, there is no viable substitute cost comparison—human attorneys remain the only option, making AI effectively unusable for this function.
Technical feasibility todayclaude-haiku-4-5-202510011/5No deployed AI system reliably performs courtroom litigation presentation end-to-end. While AI tools assist with document review and brief drafting, the live adversarial presentation of evidence in litigation remains firmly human-lawyer territory in all jurisdictions.
Technical feasibility todayclaude-sonnet-51/5No deployed product argues cases or presents evidence in court; AI tools are limited to research-stage support like drafting or document review, not live litigation advocacy.

Act as agent, trustee, guardian, or executor for businesses or individuals.

0

CI 0–0 · exposure 0 · augmentation 38 · importance 3.7/5 · click for rater detail

Panel ratings per dimension for this task
DimensionModelRatingRationale
Sector adoption velocityclaude-haiku-4-5-202510011/5No meaningful adoption of AI as independent fiduciary agents has occurred because the legal and regulatory framework prohibits it. Any adoption would require legislative change, making velocity near zero.
Sector adoption velocityclaude-sonnet-51/5This specific fiduciary function shows no meaningful AI adoption trend since it is a legally constituted role rather than a task amenable to software substitution.
Augmentation potentialclaude-haiku-4-5-202510012/5AI could assist lawyers managing fiduciary tasks by drafting documents, organizing records, or analyzing beneficiary data, but the core fiduciary decisions and execution remain human responsibilities. Assistance is limited to periphery work.
Augmentation potentialclaude-sonnet-53/5AI can assist lawyers acting as fiduciaries by drafting documents, tracking deadlines, summarizing estate assets, or flagging compliance issues, improving efficiency in the administrative aspects of the role.
Task automatabilityclaude-haiku-4-5-202510011/5This task requires fiduciary decision-making, legal judgment, and execution of decisions on behalf of others in complex, context-dependent situations. Current AI cannot independently perform fiduciary duties or make binding legal decisions with the accountability and judgment required.
Task automatabilityclaude-sonnet-51/5Serving as agent, trustee, guardian, or executor requires legal appointment, fiduciary responsibility, and personal accountability that cannot be delegated to software; AI cannot hold legal title or fiduciary duty end-to-end.
Adoption barriersclaude-haiku-4-5-202510015/5Hard legal barriers exist: only humans with proper licensing, authority, and fiduciary credentials can legally serve as agents, trustees, guardians, or executors. These roles are explicitly regulated and require human sign-off and personal liability.
Adoption barriersclaude-sonnet-55/5Fiduciary roles like trustee, guardian, and executor are defined by statute and case law requiring a legally competent person or entity, with strict liability and court oversight, making substitution essentially barred.
Cost vs. human wageclaude-haiku-4-5-202510011/5The task involves personal liability and fiduciary responsibility that cannot be offloaded to AI at any cost. Human lawyers or fiduciaries must remain accountable, making cost comparison moot—substitution is not viable.
Cost vs. human wageclaude-sonnet-51/5Since AI cannot legally assume the role at all, there is no valid cost comparison—the human fiduciary remains mandatory regardless of AI cost efficiency for supporting tasks.
Technical feasibility todayclaude-haiku-4-5-202510011/5No deployed product reliably performs fiduciary agency, trusteeship, guardianship, or executor roles. These functions require legal authority, accountability, and judgment that AI systems do not possess in production settings.
Technical feasibility todayclaude-sonnet-51/5No deployed product acts as a legally recognized fiduciary; this role is inherently vested in a natural or legal person, not an AI system.

Related occupations — Legal

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.