Judicial Law Clerks
23-1012.00Median wage $64,920/yr13,290 employed (US)Rank #371 of 923 scored · top 40% by substitution
Assist judges in court or by conducting research or preparing legal documents.
Sub-scores
0–100 · band = confidence interval from rater disagreement
Substitution31
Exposure32
Augmentation62
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
18 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%33
panel mean rating 2.3/5 → substitution pressure 33/100
Technical feasibility todayw 20%29
panel mean rating 2.2/5 → substitution pressure 29/100
Cost vs. human wagew 15%43
panel mean rating 2.7/5 → substitution pressure 43/100
Adoption barriersw 20%inverted — strong barriers lower the score25
panel mean rating 4.0/5 (barrier strength) → substitution pressure 25/100
Sector adoption velocityw 10%20
panel mean rating 1.8/5 → substitution pressure 20/100
Task breakdown (18 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.
Enter information into computerized court calendar, filing, or case management systems.
64CI 60–69 · exposure 70 · augmentation 75 · importance 3.6/5 · click for rater detail
Enter information into computerized court calendar, filing, or case management systems.
64| Dimension | Model | Rating | Rationale |
|---|---|---|---|
| Sector adoption velocity | claude-haiku-4-5-20251001 | 2/5 | Court systems are slow-moving, underfunded, and conservative in technology adoption compared to private legal and financial sectors. While some larger courts pilot automation, most still rely on manual entry by staff. |
| Sector adoption velocity | claude-sonnet-5 | 2/5 | Courts and judicial administration are traditionally slow adopters of new technology due to budget constraints, procurement rules, and cautious approach to legal-system-critical software. |
| Augmentation potential | claude-haiku-4-5-20251001 | 4/5 | AI can assist clerks by pre-populating fields, flagging inconsistencies, and organizing documents before clerk review, substantially raising productivity even where human sign-off remains required. |
| Augmentation potential | claude-sonnet-5 | 4/5 | AI-assisted form recognition, autofill, and validation tools can meaningfully speed up clerks' data entry and reduce errors while a human still verifies and finalizes entries. |
| Task automatability | claude-haiku-4-5-20251001 | 4/5 | Data entry into structured systems is highly repetitive and rule-based. Current AI can reliably extract case information from documents and populate fields in case management systems with minimal human intervention, achieving well over 50% time savings at comparable quality. |
| Task automatability | claude-sonnet-5 | 4/5 | Structured data entry into case management systems is a routine, rules-based task that AI/automation (OCR, form parsing, RPA) can handle with substantial time savings, though some fields require judgment about categorization. |
| Adoption barriers | claude-haiku-4-5-20251001 | 3/5 | Court systems often have rigid IT procurement rules, legacy system constraints, and institutional resistance to change, creating organizational friction. However, no legal mandate requires a human to perform this data entry task specifically. |
| Adoption barriers | claude-sonnet-5 | 3/5 | No licensing requirement mandates a human clerk for data entry itself, but court systems have strict procedural/audit requirements, security concerns, and legacy IT inertia that slow automation. |
| Cost vs. human wage | claude-haiku-4-5-20251001 | 5/5 | AI inference cost for document-to-system data entry is negligible (cents per case), while judicial clerk labor costs significantly more per hour, creating an order-of-magnitude cost advantage for the AI solution. |
| Cost vs. human wage | claude-sonnet-5 | 4/5 | Automated data capture and RPA tools cost far less per transaction than clerk time for repetitive entry tasks, though integration with legacy court IT systems adds overhead. |
| Technical feasibility today | claude-haiku-4-5-20251001 | 4/5 | Multiple document processing and case management integration products are deployed in courts and legal firms today (e.g., specialized legal AI platforms, RPA solutions). They reliably automate calendar and filing data entry, though some require validation and integration setup. |
| Technical feasibility today | claude-sonnet-5 | 3/5 | Court case management vendors offer e-filing intake and automated docketing tools, but many courts still rely on manual entry due to legacy systems and varied formats, so reliability in production varies widely by jurisdiction. |
Coordinate judges' meeting and appointment schedules.
62CI 55–69 · exposure 70 · augmentation 75 · importance 3.1/5 · click for rater detail
Coordinate judges' meeting and appointment schedules.
62| Dimension | Model | Rating | Rationale |
|---|---|---|---|
| Sector adoption velocity | claude-haiku-4-5-20251001 | 2/5 | Court systems are traditionally slow to adopt automation and digitize processes. While some courts use scheduling software, widespread deployment of AI-driven scheduling coordination remains limited, and many judicial chambers still rely on manual clerk-managed calendars. |
| Sector adoption velocity | claude-sonnet-5 | 2/5 | Judicial/court administration is a traditionally slow-adopting, low-digitization sector relative to finance or tech, with scheduling automation adopted unevenly and cautiously. |
| Augmentation potential | claude-haiku-4-5-20251001 | 4/5 | AI scheduling assistants can dramatically improve a clerk's productivity by handling routine conflicts, sending reminders, and proposing time slots, allowing the clerk to focus on exceptions and nuanced scheduling decisions requiring judicial input. |
| Augmentation potential | claude-sonnet-5 | 4/5 | AI calendar and scheduling assistants can meaningfully reduce back-and-forth and administrative burden for clerks managing judges' calendars, even if a human remains involved for sensitive or discretionary scheduling decisions. |
| Task automatability | claude-haiku-4-5-20251001 | 4/5 | Scheduling coordination is highly structured, involving calendar management, conflict resolution, and routine communication—tasks that current AI systems and calendar automation tools handle well. However, exceptions involving judicial discretion or complex multi-party negotiations may still require human judgment, preventing a full 5 rating. |
| Task automatability | claude-sonnet-5 | 4/5 | Scheduling coordination is a well-structured task that off-the-shelf AI scheduling assistants and calendar tools can handle end-to-end with substantial time savings, given calendar access and stated preferences. |
| Adoption barriers | claude-haiku-4-5-20251001 | 4/5 | Courts operate under formal rules and administrative procedures; judges may prefer human clerks for sensitive scheduling decisions and personal interaction. Additionally, organizational inertia and perceived need for clerk oversight of high-stakes judicial calendars create friction against full automation. |
| Adoption barriers | claude-sonnet-5 | 3/5 | No licensing requirement for scheduling itself, but judicial offices have strong confidentiality, security, and hierarchical protocol norms that create organizational friction against full automation without human oversight. |
| Cost vs. human wage | claude-haiku-4-5-20251001 | 5/5 | Automated calendar systems and AI scheduling agents cost far less per coordination task than paying a law clerk's salary ($50k–$70k annually), especially amortized across multiple judges and hundreds of scheduling events per year. |
| Cost vs. human wage | claude-sonnet-5 | 4/5 | Automated scheduling tools cost a small fraction of a clerk's salary for this narrow function, though integration with court-specific systems and oversight needs add some cost. |
| Technical feasibility today | claude-haiku-4-5-20251001 | 4/5 | Calendar management and scheduling software integrated with email/communication systems are mature and widely deployed in courts. AI-assisted scheduling tools can handle routine appointment coordination reliably at scale, though some courts may still rely on partially manual processes. |
| Technical feasibility today | claude-sonnet-5 | 3/5 | AI scheduling assistants (e.g., calendar bots, executive assistant AI tools) are deployed in many professional settings, but courts often use bespoke or legacy systems and require human judgment for prioritization, security, and confidentiality, limiting reliable production use in this specific context. |
Research laws, court decisions, documents, opinions, briefs, or other information related to cases before the court.
46CI 43–49 · exposure 50 · augmentation 88 · importance 4.8/5 · click for rater detail
Research laws, court decisions, documents, opinions, briefs, or other information related to cases before the court.
46| Dimension | Model | Rating | Rationale |
|---|---|---|---|
| Sector adoption velocity | claude-haiku-4-5-20251001 | 3/5 | Legal tech adoption is growing (major firms and some courts now use AI-assisted research tools) but remains uneven; many regional and small practices lag far behind, and courts themselves are slower adopters due to institutional inertia and conservatism around case research accuracy. |
| Sector adoption velocity | claude-sonnet-5 | 2/5 | Courts and judicial systems are traditionally slow-moving institutions with conservative technology adoption, and while some AI legal research tools are being piloted, widespread production use in judicial clerking remains limited. |
| Augmentation potential | claude-haiku-4-5-20251001 | 5/5 | Current legal AI research tools substantially enhance clerk productivity by rapidly surfacing relevant cases, generating summaries, and flagging potential precedents, allowing clerks to focus on higher-level analysis and judgment—a clear example of high augmentation even when full automation remains limited by accuracy and liability concerns. |
| Augmentation potential | claude-sonnet-5 | 4/5 | AI tools significantly speed up locating relevant cases, statutes, and precedents, and can draft preliminary summaries, substantially aiding clerks who then verify and refine the output for judicial use. |
| Task automatability | claude-haiku-4-5-20251001 | 3/5 | Current AI systems can meaningfully accelerate legal research by retrieving and summarizing relevant case law, statutes, and briefs via semantic search and document analysis. However, the full task requires nuanced legal judgment about relevance, precedential weight, and applicability to specific factual scenarios that typically demands human expertise and verification, limiting end-to-end time savings to roughly half the task. |
| Task automatability | claude-sonnet-5 | 3/5 | AI legal research tools can retrieve and summarize case law and statutes quickly, but synthesizing this into judicially relevant analysis with accurate citations and nuanced legal reasoning still requires substantial human verification, so only partial time savings are realized at equal quality today. |
| Adoption barriers | claude-haiku-4-5-20251001 | 4/5 | Significant barriers exist: courts and law firms privilege attorney oversight and malpractice liability for missed precedent or mischaracterized rulings; many judge chambers and firms have organizational resistance to full automation without attorney sign-off; and informal gatekeeping by senior attorneys slows substitution. |
| Adoption barriers | claude-sonnet-5 | 4/5 | Judicial work requires legally trained personnel, and courts have strict ethical and professional responsibility rules around accuracy and sourcing of legal research, plus documented cases of sanctions for AI-hallucinated citations, creating strong institutional caution. |
| Cost vs. human wage | claude-haiku-4-5-20251001 | 4/5 | LLM-based legal research inference is extremely cheap per query (pennies to dimes), and integration into legal workflows is straightforward, whereas a law clerk's fully loaded wage easily exceeds $60k–100k annually; the cost ratio clearly favors AI despite integration and oversight overhead. |
| Cost vs. human wage | claude-sonnet-5 | 3/5 | Subscription-based AI legal research tools cost far less than clerk hours per query, but the need for careful verification and integration into judicial workflows narrows the effective savings, keeping costs roughly comparable when factoring in oversight. |
| Technical feasibility today | claude-haiku-4-5-20251001 | 3/5 | Deployed legal AI products (ROSS Intelligence, LexisNexis+ AI-Assisted Research, Thomson Reuters Westlaw AI-Assisted Research) can reliably retrieve and summarize cases and statutes at scale, but their error rates on subtle legal distinctions and narrow precedential scope issues remain material enough that they require human review and cannot be trusted end-to-end without oversight. |
| Technical feasibility today | claude-sonnet-5 | 3/5 | Products like Westlaw Edge, Lexis+ AI, and CoCounsel are deployed in legal practice for research, but hallucination risks and need for verification mean they are not yet fully reliable for court-level legal research without attorney/clerk oversight. |
Verify that all files, complaints, or other papers are available and in the proper order.
44CI 37–51 · exposure 45 · augmentation 75 · importance 3.5/5 · click for rater detail
Verify that all files, complaints, or other papers are available and in the proper order.
44| Dimension | Model | Rating | Rationale |
|---|---|---|---|
| Sector adoption velocity | claude-haiku-4-5-20251001 | 2/5 | Legal and judicial sectors are relatively laggard in AI adoption for core processes; while some courts have digitized dockets, systematic use of AI for file verification remains limited and adoption is concentrated in larger urban courts rather than widespread. |
| Sector adoption velocity | claude-sonnet-5 | 2/5 | Courts and judicial administration are traditionally slow to adopt AI tools due to procedural rigidity, security concerns, and limited digitization in many jurisdictions. |
| Augmentation potential | claude-haiku-4-5-20251001 | 4/5 | AI-powered document checklist systems and automated flagging of missing or out-of-order documents can substantially assist law clerks by surfacing potential issues and organizing files, allowing clerks to focus on judgment calls and exceptions rather than routine scanning and ordering. |
| Augmentation potential | claude-sonnet-5 | 4/5 | AI-powered document checklists and OCR-based verification tools can meaningfully speed up a clerk's review of file completeness and ordering, even if final sign-off remains human. |
| Task automatability | claude-haiku-4-5-20251001 | 3/5 | AI can automate significant portions of file verification through document scanning, OCR, and checklist-matching against required document lists, but human judgment on 'proper order' and handling of edge cases or unusual formatting typically requires oversight, falling short of full autonomy at equal quality. |
| Task automatability | claude-sonnet-5 | 3/5 | Checking file completeness and order is a structured verification task that document management/AI systems can largely handle, though edge cases (mislabeled or ambiguous filings) still need human judgment.' |
| Adoption barriers | claude-haiku-4-5-20251001 | 4/5 | Courts operate under strict procedural rules and local court rules requiring proper filing and document ordering; automation may face resistance from judges, local bar associations, and legal precedent requiring human verification to ensure compliance and provide a clear chain of custody. |
| Adoption barriers | claude-sonnet-5 | 3/5 | While not requiring licensure itself, this task occurs within judicial proceedings where errors carry legal consequences, creating institutional caution and procedural requirements for human verification. |
| Cost vs. human wage | claude-haiku-4-5-20251001 | 4/5 | Automated document verification systems have very low per-task inference costs (scanning + classification) compared to a clerk's hourly wage for this repetitive work, though integration and initial setup costs are non-negligible. |
| Cost vs. human wage | claude-sonnet-5 | 3/5 | Automated document checking tools are cheap to run, but integration with court systems and required human oversight narrows the cost advantage over a clerk's time. |
| Technical feasibility today | claude-haiku-4-5-20251001 | 3/5 | Document management systems and AI-powered document classification tools exist and perform basic verification tasks in some courts and firms, but deployment is inconsistent and error rates on complex or non-standard document sets remain material, limiting reliable production use. |
| Technical feasibility today | claude-sonnet-5 | 2/5 | Some e-filing and document management systems flag missing or misordered documents, but no widely deployed product performs full legal-file verification reliably without human double-checking in court settings. |
Maintain judges' law libraries by assembling or updating appropriate documents.
44CI 23–65 · exposure 45 · augmentation 63 · importance 2.3/5 · click for rater detail
Maintain judges' law libraries by assembling or updating appropriate documents.
44| Dimension | Model | Rating | Rationale |
|---|---|---|---|
| Sector adoption velocity | claude-haiku-4-5-20251001 | 1/5 | The judicial sector adopts technology slowly and conservatively. Court administration remains heavily manual, and judges typically maintain personalized library systems managed by trusted clerks. Digital transformation in courts lags most other professional sectors. |
| Sector adoption velocity | claude-sonnet-5 | 2/5 | Judicial chambers and courts are traditionally slow, low-digitization environments with cautious tech adoption relative to fast-moving sectors like finance or tech-enabled professional services. |
| Augmentation potential | claude-haiku-4-5-20251001 | 3/5 | AI could usefully assist by identifying potentially relevant documents, suggesting organizational schemes, or flagging outdated materials, but human judgment about appropriateness and judicial preferences would remain essential. This represents moderate productivity enhancement within a supervised workflow. |
| Augmentation potential | claude-sonnet-5 | 4/5 | AI legal research and update-tracking tools substantially help clerks find, verify, and organize current legal materials, meaningfully boosting efficiency while human clerks retain oversight. |
| Task automatability | claude-haiku-4-5-20251001 | 2/5 | While AI can retrieve and organize legal documents from digital sources, judges' law libraries require careful curation, validation of current legal status, and understanding of judicial practice preferences. The subjective judgment about what is 'appropriate' for a specific judge and the need to ensure completeness and accuracy means substantial human review remains necessary. |
| Task automatability | claude-sonnet-5 | 4/5 | Assembling and updating legal reference materials is largely a document-management and retrieval task that current AI legal research tools can handle with significant time savings, though final verification is still needed. |
| Adoption barriers | claude-haiku-4-5-20251001 | 4/5 | Judges have specific authority over their chambers and library management; there are institutional practices and professional standards governing judicial administration. While not strictly licensed, there are court protocols and judge preferences that create meaningful organizational friction against full automation. |
| Adoption barriers | claude-sonnet-5 | 2/5 | No licensing requirement to maintain a library, though court IT policies, confidentiality practices, and clerk oversight introduce some institutional friction to full automation. |
| Cost vs. human wage | claude-haiku-4-5-20251001 | 2/5 | A law clerk performing this task is relatively inexpensive per hour compared to attorney rates. The cost of AI tools plus necessary human oversight and validation would likely approach or exceed the cost of direct clerk labor, especially given the specialized domain knowledge required. |
| Cost vs. human wage | claude-sonnet-5 | 4/5 | Automated legal database subscriptions and AI-assisted updating tools cost far less per unit of maintenance work than a clerk's time spent manually organizing and updating materials. |
| Technical feasibility today | claude-haiku-4-5-20251001 | 2/5 | No mature product reliably maintains a judge's law library end-to-end. Document management systems exist, but they require significant human oversight to verify relevance, accuracy of citations, and judicial preferences. Current AI cannot reliably determine what constitutes an 'appropriate' document for a particular judge's needs. |
| Technical feasibility today | claude-sonnet-5 | 3/5 | Legal research platforms (Westlaw, Lexis, CoCounsel) already automate citation checking, updates via annotated statutes/case law services, but full 'library maintenance' as an integrated task isn't a standalone deployed product used broadly in chambers. |
Review dockets of pending litigation to ensure adequate progress.
43CI 37–48 · exposure 45 · augmentation 75 · importance 3.5/5 · click for rater detail
Review dockets of pending litigation to ensure adequate progress.
43| Dimension | Model | Rating | Rationale |
|---|---|---|---|
| Sector adoption velocity | claude-haiku-4-5-20251001 | 2/5 | Legal organizations adopt case management tools slowly and conservatively; court systems particularly lag in AI automation due to institutional inertia, budget constraints, and judicial preference for human intermediaries. |
| Sector adoption velocity | claude-sonnet-5 | 2/5 | Court systems and judicial administration are notoriously slow to adopt new technology due to procedural conservatism, budget constraints, and data security concerns, despite AI's rapid uptake in adjacent legal services. |
| Augmentation potential | claude-haiku-4-5-20251001 | 4/5 | AI docket summary tools, status alerts, and automated case timeline visualization substantially assist clerks in monitoring multiple cases and surfacing anomalies, raising their analytical speed and coverage while preserving human judgment on adequacy. |
| Augmentation potential | claude-sonnet-5 | 4/5 | AI-powered docket management and deadline-tracking tools can meaningfully speed up a clerk's ability to spot stalled cases and prioritize review, offering strong augmentation even where full automation is unreliable. |
| Task automatability | claude-haiku-4-5-20251001 | 3/5 | AI can extract and organize case information from dockets and flag status anomalies, but reviewing for 'adequate progress' requires judgment about case complexity, court norms, and litigation strategy that current systems struggle with reliably at human expert level. |
| Task automatability | claude-sonnet-5 | 3/5 | AI can extract and summarize docket entries, flag stale cases, or identify missed deadlines, but judgment about 'adequate progress' in context often requires understanding case-specific nuance and court procedure that current systems only partially handle. |
| Adoption barriers | claude-haiku-4-5-20251001 | 4/5 | Law clerk roles are officer-of-the-court positions with implicit accountability to judges; courts typically require human review and certification of docket status, and substituting AI judgment for human legal analysis faces institutional and liability friction. |
| Adoption barriers | claude-sonnet-5 | 3/5 | This is an internal judicial administrative function without licensing requirements for the review itself, but court confidentiality, procedural rules, and institutional caution around judicial functions create moderate friction. |
| Cost vs. human wage | claude-haiku-4-5-20251001 | 3/5 | Basic docket automation (parsing, organization, alerts) is cheaper than clerk time, but the human attorney oversight requirement to validate adequacy judgments offsets significant savings; costs are roughly comparable end-to-end. |
| Cost vs. human wage | claude-sonnet-5 | 3/5 | Automated docket monitoring tools are cheap to run compared to clerk time, but integration with court systems and required human verification narrows the cost advantage to roughly comparable once oversight is included. |
| Technical feasibility today | claude-haiku-4-5-20251001 | 3/5 | Products exist for docket analytics and case management, but they typically require human verification and judgment; no deployed system reliably makes independent determinations of adequate progress without attorney oversight. |
| Technical feasibility today | claude-sonnet-5 | 2/5 | Some legal tech and case management systems offer docket tracking and deadline alerts, but reliable end-to-end review of litigation progress with legal judgment is not yet a mature deployed product in courts. |
Prepare periodic reports on court proceedings, as required.
41CI 23–60 · exposure 45 · augmentation 75 · importance 2.7/5 · click for rater detail
Prepare periodic reports on court proceedings, as required.
41| Dimension | Model | Rating | Rationale |
|---|---|---|---|
| Sector adoption velocity | claude-haiku-4-5-20251001 | 1/5 | Judicial institutions are conservative and slow to adopt automation in core functions like official record-keeping. Few courts have deployed AI for autonomous report generation; adoption remains in pilot phases if present at all. |
| Sector adoption velocity | claude-sonnet-5 | 2/5 | Judicial administration and courts are generally slow, conservative adopters of AI tools compared to fast-moving sectors like finance or tech, with pilots emerging but production use still limited. |
| Augmentation potential | claude-haiku-4-5-20251001 | 3/5 | AI can assist clerks by drafting initial summaries, flagging key procedural dates, and organizing facts, which would improve clerk productivity. However, the clerk must review, verify legal details, and ensure accuracy before submission, making AI a useful but limited assistant. |
| Augmentation potential | claude-sonnet-5 | 5/5 | AI can substantially speed up drafting, summarizing, and formatting periodic reports while the clerk retains responsibility for accuracy and final review, making this a strong augmentation use case. |
| Task automatability | claude-haiku-4-5-20251001 | 2/5 | Current AI can draft summaries of court proceedings and format routine procedural reports, but cannot reliably capture legal nuance, judicial reasoning, or discretionary determinations required for official court records. Significant human review and revision would be needed, falling short of the 50% time-savings bar. |
| Task automatability | claude-sonnet-5 | 4/5 | Summarizing and drafting periodic reports on court proceedings from transcripts, dockets, or notes is largely a language synthesis task well-suited to current LLMs, especially with structured input data.report generation with templated formats meets the time-saving bar for most of the drafting work. |
| Adoption barriers | claude-haiku-4-5-20251001 | 4/5 | Courts operate under strict procedural and evidentiary rules; periodic reports become part of official judicial records and may carry legal weight. Judges and court administrators typically require human clerk authentication and discretionary judgment, creating organizational and procedural barriers to full automation. |
| Adoption barriers | claude-sonnet-5 | 3/5 | Court reports may need clerk or judge sign-off for accuracy and confidentiality, and court systems have specific formatting/procedural rules, creating moderate institutional friction even without hard licensing requirements for this specific administrative task. |
| Cost vs. human wage | claude-haiku-4-5-20251001 | 2/5 | AI-assisted drafting may reduce some drafting time, but the cost of integration, legal oversight, and error-checking by qualified clerks keeps total cost comparable to or exceeding traditional human preparation of these reports. |
| Cost vs. human wage | claude-sonnet-5 | 4/5 | Once integrated with court records systems, AI drafting of standardized reports costs a small fraction of a clerk's billable time, though oversight and correction add some cost. |
| Technical feasibility today | claude-haiku-4-5-20251001 | 2/5 | While AI tools can summarize documents and extract procedural facts, no deployed product reliably produces court-ready periodic reports without substantial human oversight. Existing systems lack the legal precision and contextual judgment demanded by judicial institutions. |
| Technical feasibility today | claude-sonnet-5 | 3/5 | Legal AI drafting and summarization tools exist and are used in some court and law-firm settings, but reliability on legal accuracy and proper citation still requires human review, so deployment is narrower than fully autonomous reporting. |
Review complaints, petitions, motions, or pleadings that have been filed to determine issues involved or basis for relief.
36CI 25–46 · exposure 38 · augmentation 75 · importance 4.6/5 · click for rater detail
Review complaints, petitions, motions, or pleadings that have been filed to determine issues involved or basis for relief.
36| Dimension | Model | Rating | Rationale |
|---|---|---|---|
| Sector adoption velocity | claude-haiku-4-5-20251001 | 2/5 | Judicial and legal sectors are relatively slow to adopt AI automation due to liability concerns, professional licensing requirements, and institutional conservatism; while document review tools see some use, core issue-spotting remains predominantly human-performed in production court systems. |
| Sector adoption velocity | claude-sonnet-5 | 2/5 | Judiciary is a traditionally slow-adopting, highly regulated sector; while some legal AI pilots exist, court systems generally lag behind private-sector professional services in production AI deployment. |
| Augmentation potential | claude-haiku-4-5-20251001 | 4/5 | AI can significantly assist clerks by pre-organizing documents, highlighting relevant passages, flagging potential legal issues for review, and summarizing arguments, substantially raising clerk productivity while the human retains final judgment authority. |
| Augmentation potential | claude-sonnet-5 | 4/5 | AI tools can significantly speed up initial review, summarization, and issue extraction from filings, meaningfully augmenting clerks' efficiency while they retain responsibility for final analysis. |
| Task automatability | claude-haiku-4-5-20251001 | 2/5 | While AI can extract and summarize legal documents and identify formal components, determining legal issues and evaluating the basis for relief requires nuanced interpretation of precedent, jurisdiction-specific law, and discretionary judgment that current systems cannot reliably do end-to-end without substantial human review. |
| Task automatability | claude-sonnet-5 | 3/5 | AI can summarize and flag issues in legal filings quickly, but reliably identifying the precise legal basis for relief and nuanced procedural posture still requires human legal judgment and verification, limiting full end-to-end automation. |
| Adoption barriers | claude-haiku-4-5-20251001 | 4/5 | Judicial systems have strong institutional and regulatory barriers: judges must take responsibility for case determinations, attorney malpractice liability attaches to erroneous advice, and courts typically require licensed or supervised personnel to perform substantive legal analysis on filings. |
| Adoption barriers | claude-sonnet-5 | 4/5 | Judicial work carries strong barriers: clerks work under direct supervision of judges, and legal analysis feeding into judicial decisions typically requires human accountability and is subject to court rules and confidentiality constraints. |
| Cost vs. human wage | claude-haiku-4-5-20251001 | 2/5 | AI-assisted document review is cheaper per document than manual labor, but oversight, error correction, and liability management add substantial cost; the all-in cost for reliable legal issue determination remains comparable to or higher than employing law clerks. |
| Cost vs. human wage | claude-sonnet-5 | 4/5 | AI-assisted document review and summarization is dramatically cheaper per document than clerk hours, though oversight and verification by the clerk/judge add some cost back. |
| Technical feasibility today | claude-haiku-4-5-20251001 | 2/5 | AI systems can assist with document classification and keyword extraction, but no deployed product reliably performs independent issue-spotting and legal basis evaluation at the quality required for judicial use; products exist for limited document review tasks but with significant error rates in complex legal reasoning. |
| Technical feasibility today | claude-sonnet-5 | 3/5 | Legal AI tools (e.g., CoCounsel, Lexis+ AI, Harvey) are deployed in law firms and some court systems for document review and issue-spotting, but accuracy on complex or novel filings remains inconsistent, especially in high-stakes judicial contexts. |
Keep abreast of changes in the law and inform judges when cases are affected by such changes.
35CI 29–41 · exposure 30 · augmentation 75 · importance 4.1/5 · click for rater detail
Keep abreast of changes in the law and inform judges when cases are affected by such changes.
35| Dimension | Model | Rating | Rationale |
|---|---|---|---|
| Sector adoption velocity | claude-haiku-4-5-20251001 | 3/5 | Legal tech adoption is rising in corporate and large law firm settings, but courts and judicial chambers move slowly due to institutional conservatism and budget constraints; AI-assisted legal monitoring is in pilots and early adoption phases rather than production-wide deployment. |
| Sector adoption velocity | claude-sonnet-5 | 2/5 | Judicial systems and courts are generally slow adopters of AI due to conservatism, liability concerns, and lack of standardized tools tailored to case management workflows. |
| Augmentation potential | claude-haiku-4-5-20251001 | 4/5 | AI legal research tools and automated change-monitoring systems substantially assist clerks by filtering and prioritizing relevant legal updates, dramatically reducing manual review time while allowing the clerk to apply judgment and context to inform the judge. |
| Augmentation potential | claude-sonnet-5 | 4/5 | AI legal research and alert tools substantially speed up the process of identifying relevant new case law and legislative changes, meaningfully augmenting a clerk's ability to stay current while the clerk retains responsibility for judgment and communication. |
| Task automatability | claude-haiku-4-5-20251001 | 2/5 | While AI can monitor legal databases and flag statutory/regulatory changes, the task requires nuanced judgment about whether a specific case is meaningfully affected—a determination that often depends on context, prior rulings, and legal interpretation that AI struggles to reliably assess independently, preventing the 50% time-saving threshold for end-to-end automation. |
| Task automatability | claude-sonnet-5 | 2/5 | AI legal research tools can surface relevant new case law and statutory changes, but reliably tracking which pending cases are affected and synthesizing the implications requires judgment and verification that current systems cannot fully replace end-to-end.' |
| Adoption barriers | claude-haiku-4-5-20251001 | 4/5 | Judicial systems operate under strict rules of procedure and ethical obligations; courts require human professionals (attorneys, law clerks) to certify legal research and advise judges, and liability exposure for missed changes creates strong organizational and regulatory barriers to full automation. |
| Adoption barriers | claude-sonnet-5 | 3/5 | No formal licensing bars AI from doing legal research, but courts require human clerks to verify and communicate legal changes to judges, and error costs (missing a controlling change in law) are high, creating moderate institutional friction. |
| Cost vs. human wage | claude-haiku-4-5-20251001 | 3/5 | Legal research platforms and monitoring tools cost hundreds to thousands per month, comparable to or sometimes exceeding the incremental cost of a clerk spending time on this task, especially when factoring in integration and quality oversight. |
| Cost vs. human wage | claude-sonnet-5 | 3/5 | AI legal research subscriptions are relatively cheap compared to clerk salaries, but the need for human verification of accuracy in legal contexts narrows the effective cost savings, making it roughly comparable once oversight is included. |
| Technical feasibility today | claude-haiku-4-5-20251001 | 3/5 | Legal research and document monitoring tools (LexisNexis, Westlaw, AI-assisted research platforms) exist and are widely deployed, but they generate alerts and summaries requiring human review and judgment; no current system reliably determines case relevance without attorney or clerk oversight. |
| Technical feasibility today | claude-sonnet-5 | 2/5 | Legal research platforms (Westlaw AI, Lexis+ AI) can flag recent decisions and statutory amendments, but they are prone to hallucination and gaps, and no deployed product autonomously monitors and correctly maps legal changes onto specific active cases in production. |
Prepare briefs, legal memoranda, or statements of issues involved in cases, including appropriate suggestions or recommendations.
33CI 25–41 · exposure 33 · augmentation 75 · importance 4.8/5 · click for rater detail
Prepare briefs, legal memoranda, or statements of issues involved in cases, including appropriate suggestions or recommendations.
33| Dimension | Model | Rating | Rationale |
|---|---|---|---|
| Sector adoption velocity | claude-haiku-4-5-20251001 | 2/5 | Law firms and legal departments are experimenting with AI drafting assistants but remain cautious due to malpractice risk, regulatory oversight (bar rules, court rules), and liability sensitivity. Adoption is in the early, pilot phase within most organizations; large-scale production automation is rare and limited to routine document assembly, not complex brief preparation. |
| Sector adoption velocity | claude-sonnet-5 | 2/5 | Judiciary and legal clerking is a traditionally slow-adopting, conservative sector with cautious, uneven AI integration due to confidentiality, accuracy, and professional responsibility concerns. |
| Augmentation potential | claude-haiku-4-5-20251001 | 4/5 | AI legal research and drafting assistants demonstrably raise productivity when integrated as assistive tools: law clerks and attorneys use them to accelerate initial drafts, organize case law, and structure memoranda, while the human retains full analytical and review responsibility. This augmentation dynamic is already visible in practice. |
| Augmentation potential | claude-sonnet-5 | 4/5 | AI substantially speeds up legal research, issue-spotting, and first-draft generation for memos and briefs, meaningfully boosting clerk productivity while the clerk retains responsibility for accuracy and final judgment. |
| Task automatability | claude-haiku-4-5-20251001 | 2/5 | Large language models can draft sections of legal memoranda and summarize case issues, but cannot reliably perform the complete end-to-end task at equal quality. Legal briefs require nuanced judgment about argumentation strategy, precedent application, and jurisdiction-specific requirements that exceed current AI capability without substantial human revision and oversight. |
| Task automatability | claude-sonnet-5 | 3/5 | AI can draft legal memoranda and issue summaries by synthesizing case law and facts, but recommendations requiring nuanced judicial judgment and verified legal reasoning still need substantial human revision, so full end-to-end automation at equal quality isn't yet reliable. |
| Adoption barriers | claude-haiku-4-5-20251001 | 4/5 | Judicial filings and legal memoranda carry significant liability and professional responsibility: courts, bar associations, and malpractice frameworks hold human attorneys accountable for errors and fraud. The attorney must ultimately sign and take responsibility, creating a hard requirement that a qualified human legally review and authorize the output, preventing full substitution. |
| Adoption barriers | claude-sonnet-5 | 4/5 | Judicial work product requires trusted, accountable human authorship; courts have ethical and confidentiality rules, and any AI-drafted output must be reviewed and adopted by a licensed clerk or judge before use. |
| Cost vs. human wage | claude-haiku-4-5-20251001 | 2/5 | AI inference costs for legal document generation are low, but the total cost including integration, quality oversight, and human revision often rivals or exceeds the cost of having a law clerk or junior attorney draft from scratch. The requirement for expert human review negates cost advantage. |
| Cost vs. human wage | claude-sonnet-5 | 4/5 | AI-assisted drafting is dramatically cheaper per page than clerk hours, though the necessary human review and fact-checking of citations adds back some cost, keeping it below a full order-of-magnitude in effective net savings. |
| Technical feasibility today | claude-haiku-4-5-20251001 | 2/5 | Products exist (e.g., legal research AI, document drafting assistants) but they operate primarily as augmentation tools requiring heavy attorney review and refinement. No deployed system reliably produces court-ready briefs or fully formed legal memoranda meeting professional standards without material human intervention and error-checking. |
| Technical feasibility today | claude-sonnet-5 | 2/5 | Legal AI drafting tools exist and are used for research and first-draft memos, but hallucination risks and accuracy concerns mean courts and clerks still heavily verify and rewrite output rather than relying on it in production as final work product. |
Draft or proofread judicial opinions, decisions, or citations.
33CI 25–41 · exposure 33 · augmentation 63 · importance 4.8/5 · click for rater detail
Draft or proofread judicial opinions, decisions, or citations.
33| Dimension | Model | Rating | Rationale |
|---|---|---|---|
| Sector adoption velocity | claude-haiku-4-5-20251001 | 2/5 | Judicial and legal sectors have historically been conservative adopters of automation. While law firms and courts are exploring AI research and drafting tools, integration into judicial opinion writing remains rare and tentative, with heavy reliance on human verification and approval. |
| Sector adoption velocity | claude-sonnet-5 | 2/5 | Judiciary and legal clerking are historically slow to adopt new technology due to conservative institutional culture, ethics rules, and high-profile AI citation errors that have created caution. |
| Augmentation potential | claude-haiku-4-5-20251001 | 3/5 | AI can meaningfully assist clerks by drafting sections, suggesting citations, and proofreading for consistency, improving productivity on routine portions. However, core opinion-writing and legal analysis remain human-driven, limiting the transformative upside of augmentation. |
| Augmentation potential | claude-sonnet-5 | 4/5 | AI tools meaningfully speed up drafting, formatting, and citation-checking tasks for clerks who remain fully in the loop verifying legal accuracy and reasoning. |
| Task automatability | claude-haiku-4-5-20251001 | 2/5 | While AI can draft partial opinion text and check citations, judicial opinions require substantive legal reasoning, precedent synthesis, and authoritative judgment that AI cannot reliably produce. Current systems cannot meet the ≥50% time-saving bar because human review and rewriting of AI drafts is extensive, and citations must be verified against authoritative sources. |
| Task automatability | claude-sonnet-5 | 3/5 | AI can draft opinion language, summarize case facts, and check citation format quickly, but final legal reasoning, judgment application, and accuracy verification still require substantial human review, limiting full time savings at equal quality. |
| Adoption barriers | claude-haiku-4-5-20251001 | 4/5 | Judicial opinions must be authored or reviewed by a judge or licensed attorney, creating a strong legal and ethical barrier. The task carries high error-cost asymmetry: citation errors or flawed reasoning in an opinion can affect case law and litigant rights, making delegation to unaided AI infeasible regardless of capability. |
| Adoption barriers | claude-sonnet-5 | 4/5 | Judicial opinions require signoff by judges and clerks under strict professional and ethical rules, with high liability for errors and citation fabrication, creating strong institutional resistance to full automation. |
| Cost vs. human wage | claude-haiku-4-5-20251001 | 2/5 | AI research and drafting tools carry meaningful licensing and integration costs. When accounting for required human oversight, verification, and rework, the all-in cost remains comparable to or higher than paying a clerk for original work, especially given the high stakes of judicial accuracy. |
| Cost vs. human wage | claude-sonnet-5 | 4/5 | AI drafting and citation-checking tools are far cheaper per unit of output than clerk hours, though human oversight costs remain necessary and reduce the ratio somewhat. |
| Technical feasibility today | claude-haiku-4-5-20251001 | 2/5 | AI tools can assist with citation formatting and basic proofreading, but no deployed product reliably drafts full judicial opinions or makes binding citation decisions. Products exist for legal research and draft support, but they require heavy human oversight and cannot replace the core intellectual work. |
| Technical feasibility today | claude-sonnet-5 | 2/5 | Legal AI drafting/proofreading tools exist (e.g., citation checkers, legal LLM assistants) but are not yet reliably deployed for judicial opinion drafting given hallucination risks and the sensitivity of judicial work. |
Respond to questions from judicial officers or court staff on general legal issues.
33CI 29–37 · exposure 30 · augmentation 75 · importance 3.3/5 · click for rater detail
Respond to questions from judicial officers or court staff on general legal issues.
33| Dimension | Model | Rating | Rationale |
|---|---|---|---|
| Sector adoption velocity | claude-haiku-4-5-20251001 | 2/5 | Court systems are traditionally slow adopters of automation due to budget constraints, regulatory conservatism, and institutional inertia. Pilots of legal AI in courts exist but production deployment at scale remains limited; the judiciary lags finance and professional services sectors significantly. |
| Sector adoption velocity | claude-sonnet-5 | 2/5 | The judiciary and court systems are historically slow to adopt AI due to conservatism, ethical rules, and liability concerns, with pilots emerging but production use for direct judicial questions still rare. |
| Augmentation potential | claude-haiku-4-5-20251001 | 4/5 | AI can substantially assist clerks by drafting initial legal memoranda, summarizing case law, and organizing statutes, allowing the clerk to focus on judgment and context-specific refinement. This augmentation significantly raises clerk productivity while the clerk retains final responsibility. |
| Augmentation potential | claude-sonnet-5 | 4/5 | AI can significantly speed up legal research, case law lookup, and drafting of preliminary answers, meaningfully augmenting a clerk's ability to respond to legal questions while the clerk retains final judgment. |
| Task automatability | claude-haiku-4-5-20251001 | 2/5 | While AI can draft initial responses to general legal questions, judicial officers require nuanced, context-aware advice that accounts for specific jurisdictional details, recent case law, and the judge's particular bench philosophy. Current AI systems cannot reliably handle the full task end-to-end with 50% time savings at equal quality; a human clerk must verify and adapt responses. |
| Task automatability | claude-sonnet-5 | 2/5 | Some general legal research and question-answering could be assisted by AI, but the task requires nuanced judgment, context about a specific case, and interactive dialogue that current systems cannot fully replace end-to-end. |
| Adoption barriers | claude-haiku-4-5-20251001 | 4/5 | Judges and court systems have strong preferences for human accountability and verification of legal advice. Liability exposure, precedent reliance, and organizational culture within courts create friction against full automation. Many jurisdictions also require human certification for certain legal advisory tasks. |
| Adoption barriers | claude-sonnet-5 | 4/5 | Judicial answers touch on legal interpretation and procedural accuracy where errors carry high liability; courts typically require a qualified human (often bar-admitted) to verify or provide such guidance, creating strong professional and ethical barriers. |
| Cost vs. human wage | claude-haiku-4-5-20251001 | 4/5 | AI inference and integration costs for legal question-answering are relatively low, particularly compared to the loaded wage of a law clerk ($50–80k+). Even accounting for oversight and human verification, AI assistance is substantially cheaper than hiring additional clerical staff. |
| Cost vs. human wage | claude-sonnet-5 | 3/5 | AI legal research tools are cheap per query, but given the need for human verification, oversight, and liability concerns, the effective cost is roughly comparable to using a trained clerk for this function. |
| Technical feasibility today | claude-haiku-4-5-20251001 | 3/5 | LLMs and legal AI tools can generate initial answers to general legal questions and are deployed in some legal tech platforms, but they produce material errors in jurisdiction-specific law, procedural nuance, and citation accuracy. No mature product reliably substitutes for a human clerk's judgment in responding to judicial officers. |
| Technical feasibility today | claude-sonnet-5 | 2/5 | Legal research assistants and LLM-based tools exist and are used for drafting and research support, but no deployed product reliably fields real-time legal questions from judges or court staff with the accuracy and trust required in a courtroom setting. |
Communicate with counsel regarding case management or procedural requirements.
24CI 23–25 · exposure 25 · augmentation 50 · importance 3.4/5 · click for rater detail
Communicate with counsel regarding case management or procedural requirements.
24| Dimension | Model | Rating | Rationale |
|---|---|---|---|
| Sector adoption velocity | claude-haiku-4-5-20251001 | 1/5 | Judicial institutions are slow to adopt technology, highly risk-averse, and bound by centuries of procedural tradition and ethical rules. Counsel communication is a core function of the judge's office, not a candidate for automation in practice. |
| Sector adoption velocity | claude-sonnet-5 | 2/5 | Courts and judicial administration are traditionally slow to adopt AI tools due to procedural formality, confidentiality concerns, and conservative institutional culture, resulting in limited production deployment. |
| Augmentation potential | claude-haiku-4-5-20251001 | 3/5 | AI can assist by drafting routine procedural notices and organizing case deadlines, reducing clerk cognitive load on routine tasks. However, the human clerk must review, contextualize, and take responsibility for each communication, limiting the augmentation benefit. |
| Augmentation potential | claude-sonnet-5 | 3/5 | AI can help clerks draft communications, summarize case status, and track procedural deadlines, offering useful support even though the actual interaction and judgment remain human-led. |
| Task automatability | claude-haiku-4-5-20251001 | 2/5 | While AI can draft routine communications about procedural matters, the nuanced, context-dependent communication with counsel—often involving case strategy, informal negotiation, and relationship-building—requires human judgment and professional accountability. Only simple, templated procedural notifications could be fully automated. |
| Task automatability | claude-sonnet-5 | 2/5 | This task involves nuanced, context-sensitive judgment about procedural posture, court preferences, and case-specific dynamics that require real-time interactive communication; AI can draft or summarize but cannot reliably conduct these exchanges end-to-end today. |
| Adoption barriers | claude-haiku-4-5-20251001 | 4/5 | Courts and judicial ethics rules require that communications from chambers be made by or under direct supervision of the judge or authorized staff; liability for procedural missteps falls on the judge. Counsel expect verified human accountability, and rules of professional conduct restrict who can communicate on behalf of the court. |
| Adoption barriers | claude-sonnet-5 | 4/5 | Judicial clerks operate under court authority and confidentiality/ethical rules; communications with counsel about procedure often carry quasi-official weight requiring appropriate authorization and accountability, creating meaningful institutional and professional-responsibility barriers. |
| Cost vs. human wage | claude-haiku-4-5-20251001 | 2/5 | A judicial law clerk's time on these communications (20–30 min per matter) is relatively low-cost; AI-assisted drafting might save 10–15 min per communication after oversight, making the all-in cost comparable or slightly higher given integration overhead. |
| Cost vs. human wage | claude-sonnet-5 | 2/5 | While AI drafting assistance is cheap, the human clerk still must review, verify accuracy against court rules, and handle the actual authorized communication, so overall cost savings are modest given oversight needs. |
| Technical feasibility today | claude-haiku-4-5-20251001 | 2/5 | No deployed legal system relies on AI to independently conduct counsel communications; this remains a human responsibility. Some contract-drafting and templating tools exist, but actual counsel communication requires human professional judgment and signature authority. |
| Technical feasibility today | claude-sonnet-5 | 2/5 | Some legal AI tools assist with drafting scheduling communications or summarizing procedural status, but no deployed product autonomously manages attorney-clerk communications in production court settings. |
Supervise law students, volunteers, or other personnel assigned to the court.
3CI 0–5 · exposure 0 · augmentation 25 · importance 2.6/5 · click for rater detail
Supervise law students, volunteers, or other personnel assigned to the court.
3| Dimension | Model | Rating | Rationale |
|---|---|---|---|
| Sector adoption velocity | claude-haiku-4-5-20251001 | 1/5 | This task is embedded in judicial governance and personnel management, sectors that have shown minimal AI adoption for core management functions and require human authority and accountability. |
| Sector adoption velocity | claude-sonnet-5 | 1/5 | Court systems are slow to adopt AI generally, and personnel supervision specifically sees essentially no AI-driven displacement or tooling adoption. |
| Augmentation potential | claude-haiku-4-5-20251001 | 2/5 | AI might assist with scheduling or performance documentation compilation, but supervision—evaluating competence, providing feedback, making personnel decisions—requires direct human judgment and cannot be meaningfully augmented by AI systems. |
| Augmentation potential | claude-sonnet-5 | 2/5 | AI could help with administrative aspects like scheduling or tracking assignments, but offers minimal assistance for the interpersonal and evaluative core of supervision. |
| Task automatability | claude-haiku-4-5-20251001 | 1/5 | Supervising personnel requires real-time judgment, relationship management, performance evaluation, and accountability—human skills that AI cannot replicate. Current AI systems cannot assess competence, provide mentorship, or make personnel decisions. |
| Task automatability | claude-sonnet-5 | 1/5 | Supervision involves personnel management, mentoring, and real-time judgment about people's performance and development, which current AI cannot perform end-to-end. |
| Adoption barriers | claude-haiku-4-5-20251001 | 5/5 | Judicial courts operate under strict hierarchical authority and legal accountability structures. A licensed or appointed judicial officer must legally supervise court personnel; this is a structural requirement of the judicial system. |
| Adoption barriers | claude-sonnet-5 | 4/5 | Supervisory and mentorship roles carry organizational accountability and often implicit authority structures that require a human supervisor, though not a formal licensing requirement specifically for supervision. |
| Cost vs. human wage | claude-haiku-4-5-20251001 | 1/5 | Supervision is inherently a human function with legal accountability; AI deployment would not reduce the need for a human supervisor, making the cost comparison non-applicable and unfavorable for AI substitution. |
| Cost vs. human wage | claude-sonnet-5 | 1/5 | There is no viable AI substitute for supervisory responsibility, so cost comparison favors the human entirely. |
| Technical feasibility today | claude-haiku-4-5-20251001 | 1/5 | No deployed AI system can autonomously supervise human workers or make personnel management decisions. This task requires human authority, legal responsibility, and interpersonal dynamics that exceed current AI capabilities. |
| Technical feasibility today | claude-sonnet-5 | 1/5 | No deployed product manages or supervises human personnel in a legal/court setting; this remains firmly a human management function. |
Confer with judges concerning legal questions, construction of documents, or granting of orders.
2CI 0–4 · exposure 0 · augmentation 38 · importance 4.6/5 · click for rater detail
Confer with judges concerning legal questions, construction of documents, or granting of orders.
2| Dimension | Model | Rating | Rationale |
|---|---|---|---|
| Sector adoption velocity | claude-haiku-4-5-20251001 | 1/5 | Court systems remain highly conservative in automation, with strong institutional preference for human clerks who understand chambers workflow and judicial discretion; adoption of AI for judge conferral is negligible. |
| Sector adoption velocity | claude-sonnet-5 | 1/5 | Courts are notoriously slow to adopt AI for substantive judicial functions due to confidentiality, due process, and professional responsibility concerns. |
| Augmentation potential | claude-haiku-4-5-20251001 | 2/5 | AI can assist a clerk in preparing research summaries or document drafts before a conference with the judge, but it offers limited assistance during the actual conferral itself, which depends on real-time legal reasoning and interpersonal communication. |
| Augmentation potential | claude-sonnet-5 | 3/5 | AI can help clerks research case law, draft summaries, or prepare talking points before conferring with judges, improving preparation without touching the actual judge-clerk conference itself. |
| Task automatability | claude-haiku-4-5-20251001 | 1/5 | This task requires real-time dialogue with a judge on complex legal and procedural matters, where the clerk must understand the judge's intent, priorities, and prior decisions. Current AI systems cannot reliably participate in consequential legal deliberation or adapt to a specific judge's reasoning in a way that would meet the 50% time-saving bar while maintaining equal quality. |
| Task automatability | claude-sonnet-5 | 1/5 | This is a live, interactive deliberative conference requiring real-time judgment exchange and trust between a specific judge and clerk; AI cannot substitute for this interpersonal advisory dialogue today.' |
| Adoption barriers | claude-haiku-4-5-20251001 | 5/5 | The task involves direct legal and procedural guidance from a judge; delegation to an AI system without human intermediation would likely violate judicial practice norms and create liability concerns, as the judge relies on the clerk's professional judgment and legal training. |
| Adoption barriers | claude-sonnet-5 | 5/5 | Judicial deliberation and legal advice to a judge are core judicial functions with strict confidentiality, ethical, and constitutional requirements demanding a qualified human clerk/attorney. |
| Cost vs. human wage | claude-haiku-4-5-20251001 | 2/5 | AI document drafting and legal research are inexpensive at scale, but the marginal cost of a junior clerk's time spent in brief confab with a judge is low relative to integration and oversight overhead; AI does not yet eliminate that human interaction. |
| Cost vs. human wage | claude-sonnet-5 | 1/5 | There is no viable AI product performing this conferencing role, so no meaningful cost comparison exists—the human clerk is the only option. |
| Technical feasibility today | claude-haiku-4-5-20251001 | 1/5 | No deployed product reliably conducts authentic legal conferral with judges on document construction or orders. While AI can draft legal documents and answer legal questions, it cannot engage in the back-and-forth dialogue with judicial decision-makers that this task requires. |
| Technical feasibility today | claude-sonnet-5 | 1/5 | No deployed product conducts substantive legal conferencing with a judge on case-specific orders; this remains firmly human and confidential judicial work. |
Attend court sessions to hear oral arguments or record necessary case information.
1CI 0–3 · exposure 0 · augmentation 38 · importance 4.0/5 · click for rater detail
Attend court sessions to hear oral arguments or record necessary case information.
1| Dimension | Model | Rating | Rationale |
|---|---|---|---|
| Sector adoption velocity | claude-haiku-4-5-20251001 | 1/5 | This task involves mandatory physical presence in court, which has remained essentially unchanged across decades of legal practice. Adoption velocity toward AI substitution is zero because the task's requirement for human attendance is non-negotiable within the current legal system. |
| Sector adoption velocity | claude-sonnet-5 | 2/5 | Legal sector adoption of AI is growing for research and drafting, but courtroom attendance functions remain minimally touched by AI due to procedural and access constraints. |
| Augmentation potential | claude-haiku-4-5-20251001 | 2/5 | AI could marginally assist by post-processing recordings or transcripts after court, but during the session itself—the core of this task—AI offers minimal assistance. The clerk's role requires real-time attention and judgment that current AI cannot meaningfully enhance in the moment. |
| Augmentation potential | claude-sonnet-5 | 3/5 | AI can assist with transcription, summarization, and note organization from recordings, aiding the clerk's later work product even though it cannot replace live attendance. |
| Task automatability | claude-haiku-4-5-20251001 | 1/5 | Attending court sessions and hearing oral arguments requires physical presence in a legal proceeding, which current AI systems cannot do. While AI can transcribe or summarize arguments from recordings post-hoc, it cannot substitute for the presence and attention required to observe and record case information in real time during active proceedings. |
| Task automatability | claude-sonnet-5 | 1/5 | Physical/virtual attendance and real-time contextual comprehension of live courtroom proceedings, including nonverbal cues and judicial intent, cannot be end-to-end automated by current AI systems. |
| Adoption barriers | claude-haiku-4-5-20251001 | 5/5 | Legal proceedings are heavily regulated; rules of court, evidence rules, and due process protections mandate human observers and record-keepers in official capacities. A human court officer must physically attend and document proceedings, creating absolute legal and institutional barriers to automation. |
| Adoption barriers | claude-sonnet-5 | 5/5 | Court proceedings require a legally recognized human presence (clerk, attorney, or authorized officer) for confidentiality, procedural, and evidentiary integrity reasons, making this a hard institutional barrier. |
| Cost vs. human wage | claude-haiku-4-5-20251001 | 1/5 | The task inherently requires human physical presence, making AI cost comparison inapplicable. A clerk's labor is necessary regardless; AI cannot reduce the cost of this specific requirement. |
| Cost vs. human wage | claude-sonnet-5 | 1/5 | There is no viable AI substitute performing this task, so no meaningful cost comparison favors AI; any partial tooling adds cost on top of required human presence. |
| Technical feasibility today | claude-haiku-4-5-20251001 | 1/5 | No deployed AI system can physically attend court or participate as a substitute for human attendance. Courts require verified human attendance and observation of proceedings, which is outside the scope of current AI capabilities. |
| Technical feasibility today | claude-sonnet-5 | 1/5 | No deployed product attends and represents a clerk's presence in court proceedings; transcription tools exist but do not perform the substantive attending/interpreting role. |
Participate in conferences or discussions between trial attorneys and judges.
0CI 0–0 · exposure 0 · augmentation 25 · importance 2.8/5 · click for rater detail
Participate in conferences or discussions between trial attorneys and judges.
0| Dimension | Model | Rating | Rationale |
|---|---|---|---|
| Sector adoption velocity | claude-haiku-4-5-20251001 | 1/5 | There is no meaningful adoption of AI for this task because the judicial system requires human legal professionals to attend and participate in conferences; no sector pressure or technology innovation can bypass this requirement. |
| Sector adoption velocity | claude-sonnet-5 | 1/5 | Judicial and legal proceedings are among the most conservative, tradition-bound, and heavily regulated environments, with minimal AI adoption in live courtroom or chambers interactions. |
| Augmentation potential | claude-haiku-4-5-20251001 | 2/5 | AI could assist a law clerk by preparing case summaries or organizing documents before a conference, but it cannot augment the actual participation or discussion, which remains a human-only activity requiring legal judgment and presence. |
| Augmentation potential | claude-sonnet-5 | 2/5 | AI might help clerks prepare background research or notes before such conferences, but it offers no real-time assistance during the actual discussions themselves. |
| Task automatability | claude-haiku-4-5-20251001 | 1/5 | This task requires real-time interpersonal dialogue, nuanced legal judgment, and presence in confidential judicial discussions where AI cannot legally or practically participate. Current AI cannot meaningfully automate attendance or substantive participation in attorney-judge conferences. |
| Task automatability | claude-sonnet-5 | 1/5 | This requires live, in-person judgment, presence, and real-time interpersonal interaction between judges and attorneys, none of which current AI can substitute for.dev |
| Adoption barriers | claude-haiku-4-5-20251001 | 5/5 | Strong legal and regulatory barriers protect this task: only licensed attorneys and judges (or their authorized human staff) can participate in judicial conferences, and courts require human presence and accountability for confidential discussions affecting case outcomes. |
| Adoption barriers | claude-sonnet-5 | 5/5 | Court proceedings require licensed, authorized human participants (clerks, attorneys, judges) with legal standing; an AI cannot legally substitute for a human in these proceedings. |
| Cost vs. human wage | claude-haiku-4-5-20251001 | 1/5 | AI systems cannot perform this task at all, making cost comparison moot; a human law clerk remains essential and cannot be replaced by an AI system at any price point. |
| Cost vs. human wage | claude-sonnet-5 | 1/5 | There is no AI system performing this role at all, so no cost comparison favors AI; the human clerk's presence is the entire deliverable. |
| Technical feasibility today | claude-haiku-4-5-20251001 | 1/5 | No deployed product can reliably participate in judicial conferences; this would require legal authority and real-time presence that AI systems do not possess. The task is fundamentally human-centric and legally bound to human participants. |
| Technical feasibility today | claude-sonnet-5 | 1/5 | No deployed AI product participates as a stand-in in live judicial conferences or attorney-judge discussions; this is not a task any product targets. |
Perform courtroom duties, including calling calendars, administering oaths, and swearing in jury panels and witnesses.
0CI 0–0 · exposure 0 · augmentation 0 · importance 2.2/5 · click for rater detail
Perform courtroom duties, including calling calendars, administering oaths, and swearing in jury panels and witnesses.
0| Dimension | Model | Rating | Rationale |
|---|---|---|---|
| Sector adoption velocity | claude-haiku-4-5-20251001 | 1/5 | The judicial system is highly regulated and conservative with automation; courtroom procedures are governed by rules of court and statute. There is no evidence of AI deployment for these core judicial functions. |
| Sector adoption velocity | claude-sonnet-5 | 1/5 | Court systems are notoriously slow to adopt automation for procedural and ceremonial functions, and no momentum exists toward AI performing this specific task. |
| Augmentation potential | claude-haiku-4-5-20251001 | 1/5 | AI offers no meaningful assistance for calling calendars, administering oaths, or swearing in witnesses, as these are ceremonial and procedural acts that require human authority and legal standing to execute. |
| Augmentation potential | claude-sonnet-5 | 1/5 | AI offers essentially no assistance to the physical, ceremonial act of calling calendars or administering oaths in a live courtroom setting. |
| Task automatability | claude-haiku-4-5-20251001 | 1/5 | This task is inherently procedural and ceremonial, requiring physical presence in a courtroom, verbal administration of oaths with legal formality, and real-time interaction with juries and witnesses. Current AI systems cannot physically perform these in-person courtroom functions or manage the temporal, sequential coordination needed. |
| Task automatability | claude-sonnet-5 | 1/5 | This is a live, in-person ceremonial and procedural courtroom function requiring physical presence, real-time interaction, and legal authority; no AI system can perform these acts today. |
| Adoption barriers | claude-haiku-4-5-20251001 | 5/5 | This task is heavily protected by law and court procedure: only authorized court staff (judges, clerks, or designated officers) may legally administer oaths, call calendars, and conduct jury proceedings. Substitution with AI is legally prohibited regardless of technical capability. |
| Adoption barriers | claude-sonnet-5 | 5/5 | Administering oaths and swearing in witnesses/juries is a formal legal act typically requiring an authorized court officer, with procedural and evidentiary rules mandating human performance. |
| Cost vs. human wage | claude-haiku-4-5-20251001 | 1/5 | AI systems have no current mechanism to perform these duties, making cost comparison inapplicable. A human court clerk remains the only viable option. |
| Cost vs. human wage | claude-sonnet-5 | 1/5 | There is no viable AI substitute for this in-person, legally-authorized act, so cost comparison is moot; any attempt would require human oversight anyway, offering no savings. |
| Technical feasibility today | claude-haiku-4-5-20251001 | 1/5 | No deployed AI system can call courtroom calendars, administer oaths, or manage jury procedures in a live courtroom setting. These tasks require authorized court personnel with physical presence and legal standing to conduct official proceedings. |
| Technical feasibility today | claude-sonnet-5 | 1/5 | No deployed product administers oaths, swears in juries, or calls courtroom calendars in live proceedings; this remains entirely a research-irrelevant, human-performed function. |
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.