Court Reporters and Simultaneous Captioners
27-3092.00Median wage $72,420/yr12,870 employed (US)Rank #106 of 923 scored · top 11% by substitution
Use verbatim methods and equipment to capture, store, retrieve, and transcribe pretrial and trial proceedings or other information. Includes stenocaptioners who operate computerized stenographic captioning equipment to provide captions of live or prerecorded broadcasts for hearing-impaired viewers.
Sub-scores
0–100 · band = confidence interval from rater disagreement
Substitution46
Exposure50
Augmentation68
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
14 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
7%
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%50
panel mean rating 3.0/5 → substitution pressure 50/100
Technical feasibility todayw 20%50
panel mean rating 3.0/5 → substitution pressure 50/100
Cost vs. human wagew 15%57
panel mean rating 3.3/5 → substitution pressure 57/100
Adoption barriersw 20%inverted — strong barriers lower the score31
panel mean rating 3.7/5 (barrier strength) → substitution pressure 31/100
Sector adoption velocityw 10%35
panel mean rating 2.4/5 → substitution pressure 35/100
Task breakdown (14 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.
Proofread transcripts for correct spelling of words.
79CI 76–81 · exposure 80 · augmentation 88 · importance 4.8/5 · click for rater detail
Proofread transcripts for correct spelling of words.
79| Dimension | Model | Rating | Rationale |
|---|---|---|---|
| Sector adoption velocity | claude-haiku-4-5-20251001 | 3/5 | Legal and court reporting sectors have moderate digitization and adopt automated proofreading tools, but adoption is uneven; many courts and firms still rely on manual review, limiting deep production-scale displacement. |
| Sector adoption velocity | claude-sonnet-5 | 3/5 | Legal/court transcription is a moderately conservative sector; spell-check tools are used but full AI-based proofreading integration into certified legal transcript workflows is still emerging rather than deeply embedded. |
| Augmentation potential | claude-haiku-4-5-20251001 | 4/5 | Spelling-correction tools substantially enhance proofreader productivity by flagging errors for human review and correction, allowing the proofreader to focus on context-dependent and homophone issues that require judgment. |
| Augmentation potential | claude-sonnet-5 | 5/5 | AI spell-checking is a mature, widely used augmentation that significantly speeds up human proofreading of transcripts while the reporter retains final review responsibility. |
| Task automatability | claude-haiku-4-5-20251001 | 4/5 | AI can automatically detect and correct spelling errors in transcripts with high accuracy using spell-check and contextual language models, easily meeting the 50% time-saving threshold. Manual proofreading remains necessary for judgment calls, but the core spelling task is well-automated. |
| Task automatability | claude-sonnet-5 | 4/5 | Spell-checking and proofreading text against context is a well-solved NLP task; modern AI can flag and correct misspellings and homophone errors in transcripts with high accuracy and speed., |
| Adoption barriers | claude-haiku-4-5-20251001 | 2/5 | While court transcripts may have some regulatory requirements around accuracy, spelling correction itself is not legally gated to licensed humans, and existing tools are widely integrated into workflows with minimal authorization friction. |
| Adoption barriers | claude-sonnet-5 | 2/5 | Court reporting has certification/liability requirements for the overall transcript's accuracy, but the narrow task of spelling proofreading itself is not separately regulated or requiring sign-off. |
| Cost vs. human wage | claude-haiku-4-5-20251001 | 5/5 | Automated spell-checking via commodity software or LLM APIs costs pennies per transcript, orders of magnitude cheaper than the loaded wage of a human proofreader for the same volume. |
| Cost vs. human wage | claude-sonnet-5 | 5/5 | Automated spell-checking costs fractions of a cent per document compared to the loaded hourly cost of a human proofreader performing the same check. |
| Technical feasibility today | claude-haiku-4-5-20251001 | 5/5 | Deployed spell-checking and grammar-correction systems (including LLM-based tools) reliably perform spelling verification at scale in production environments across document-processing workflows. |
| Technical feasibility today | claude-sonnet-5 | 4/5 | Grammar/spell-check tools (e.g., Grammarly, built-in editor tools) and LLM-based proofreading are already deployed widely and reliably catch spelling errors in production text workflows. |
File and store shorthand notes of court session.
69CI 62–76 · exposure 75 · augmentation 63 · importance 4.7/5 · click for rater detail
File and store shorthand notes of court session.
69| Dimension | Model | Rating | Rationale |
|---|---|---|---|
| Sector adoption velocity | claude-haiku-4-5-20251001 | 3/5 | Court systems have digitized filing and case management, but adoption is uneven across jurisdictions and often constrained by legacy systems, budgets, and institutional conservatism typical of government sectors. |
| Sector adoption velocity | claude-sonnet-5 | 3/5 | Legal/court systems are historically slower to digitize due to bureaucracy and varying jurisdictional rules, though electronic filing systems have been adopted in many courts over the past decade. |
| Augmentation potential | claude-haiku-4-5-20251001 | 4/5 | AI-powered tagging, full-text search, metadata extraction, and automated filing assist court reporters and administrators significantly, allowing faster retrieval and organization of session notes while humans retain control over accuracy and completeness. |
| Augmentation potential | claude-sonnet-5 | 3/5 | AI-assisted document management tools can help organize, tag, and retrieve filed notes, improving efficiency, though this is a narrow, low-complexity administrative task. |
| Task automatability | claude-haiku-4-5-20251001 | 4/5 | Filing and storing digital shorthand notes can be largely automated via document management systems, OCR, and database indexing. However, the initial conversion of stenographic shorthand to machine-readable format and subsequent quality verification still require significant human oversight, preventing a full end-to-end 50% time-saving threshold. |
| Task automatability | claude-sonnet-5 | 4/5 | Filing and storing digital notes/transcripts is a straightforward document management task easily handled by automated storage systems, cloud repositories, and case management software with minimal human oversight needed. |
| Adoption barriers | claude-haiku-4-5-20251001 | 4/5 | Court records are subject to regulatory oversight and retention requirements; chain-of-custody rules and authentication standards create legal obligations that require human accountability, though the filing mechanism itself is not statutorily reserved to licensed professionals. |
| Adoption barriers | claude-sonnet-5 | 2/5 | Some court systems require certified custody chains and record-keeping compliance, but pure filing/storage of notes carries low liability risk and is largely administrative rather than requiring licensed judgment. |
| Cost vs. human wage | claude-haiku-4-5-20251001 | 4/5 | Digital storage and automated filing via commercial document management systems cost far less than manual filing labor, with integration costs amortized across many tasks, making AI solutions substantially cheaper than human filing and archiving. |
| Cost vs. human wage | claude-sonnet-5 | 5/5 | Automated digital storage and filing is vastly cheaper than paying a human to manually file and archive notes, requiring only minimal cloud storage costs. |
| Technical feasibility today | claude-haiku-4-5-20251001 | 4/5 | Document management and storage systems are mature and deployed across courts; however, reliable automated stenographic shorthand recognition remains imperfect and typically requires human validation, limiting production-grade autonomy. |
| Technical feasibility today | claude-sonnet-5 | 4/5 | Court and legal document management systems already automate filing, indexing, and retrieval of transcripts and notes in production environments across many jurisdictions. |
Transcribe recorded proceedings in accordance with established formats.
64CI 62–66 · exposure 75 · augmentation 100 · importance 4.8/5 · click for rater detail
Transcribe recorded proceedings in accordance with established formats.
64| Dimension | Model | Rating | Rationale |
|---|---|---|---|
| Sector adoption velocity | claude-haiku-4-5-20251001 | 3/5 | Courts and legal firms are piloting and adopting AI transcription for routine recorded proceedings, but adoption remains uneven across jurisdictions due to regulatory variation and institutional conservatism. Production adoption is increasing but not yet the dominant pattern. |
| Sector adoption velocity | claude-sonnet-5 | 3/5 | Legal and captioning industries are adopting AI transcription tools steadily but cautiously due to accuracy, certification, and evidentiary requirements, resulting in moderate but growing adoption. |
| Augmentation potential | claude-haiku-4-5-20251001 | 5/5 | AI transcription powerfully augments human court reporters by handling initial transcription and time-intensive formatting, allowing reporters to focus on accuracy verification, legal terminology, and speaker identification rather than raw transcription, dramatically raising overall productivity. |
| Augmentation potential | claude-sonnet-5 | 5/5 | AI transcription dramatically speeds up draft generation, letting reporters/captioners focus on verification, formatting, and correction rather than manual transcription from scratch. |
| Task automatability | claude-haiku-4-5-20251001 | 4/5 | Speech-to-text AI can transcribe recorded proceedings with high accuracy (>95% word error rate in clean audio), meeting the 50% time-saving threshold when combined with minimal human review. The main remaining overhead is format verification and legal-specific terminology correction, which reduces but does not eliminate the time savings. |
| Task automatability | claude-sonnet-5 | 4/5 | Modern ASR systems can transcribe recorded speech to text with high accuracy, and formatting templates can be applied programmatically, meeting the time-saving threshold for much of the task, though verification and formatting nuance still requires human review. |
| Adoption barriers | claude-haiku-4-5-20251001 | 4/5 | Court reporting rules, judicial requirements, and admissibility standards in many jurisdictions mandate human certification or oversight of transcripts; some courts legally require a certified court reporter to prepare official records. These licensing and liability barriers meaningfully slow substitution even where AI could perform the transcription task itself. |
| Adoption barriers | claude-sonnet-5 | 4/5 | Many jurisdictions require certified court reporters or legally recognized transcript preparers for official records, creating licensing and admissibility barriers that limit full replacement. |
| Cost vs. human wage | claude-haiku-4-5-20251001 | 4/5 | AI transcription costs (typically $0.10–$0.50 per minute) are substantially cheaper than court reporter labor ($30–$50/hour loaded), achieving roughly 5–10x cost advantage for recorded material, though integration and QA add overhead. |
| Cost vs. human wage | claude-sonnet-5 | 5/5 | Automated transcription costs a small fraction of a certified court reporter's hourly rate, even after factoring in human review and correction time. |
| Technical feasibility today | claude-haiku-4-5-20251001 | 4/5 | Deployed products (e.g., Otter.ai, Rev, professional speech-to-text APIs) reliably transcribe recorded proceedings at scale in production environments. Error rates on clear courtroom audio are low enough for practical use, though some legal contexts still require human verification of outputs. |
| Technical feasibility today | claude-sonnet-5 | 4/5 | Deployed ASR/transcription products (e.g., Otter.ai, Verbit, professional legal transcription platforms) are used in production for legal and captioning contexts, though certified verbatim court transcripts still typically require human correction. |
Record symbols on computer storage media and use computer aided transcription to translate and display them as text.
57CI 49–66 · exposure 62 · augmentation 75 · importance 4.8/5 · click for rater detail
Record symbols on computer storage media and use computer aided transcription to translate and display them as text.
57| Dimension | Model | Rating | Rationale |
|---|---|---|---|
| Sector adoption velocity | claude-haiku-4-5-20251001 | 3/5 | Legal and court sectors show moderate adoption of AI transcription tools, with growing pilot programs and adoption in depositions and some courtrooms, but official court proceedings still require licensed human court reporters in many jurisdictions, slowing deep replacement. |
| Sector adoption velocity | claude-sonnet-5 | 3/5 | Captioning and general transcription sectors have adopted AI quickly (media, education), but official court reporting remains slower due to certification and legal admissibility requirements. |
| Augmentation potential | claude-haiku-4-5-20251001 | 4/5 | AI transcription powerfully augments court reporters by handling real-time symbol-to-text conversion and providing instant drafts that reporters review and certify, dramatically increasing their productivity while they maintain control over accuracy and legal validity. |
| Augmentation potential | claude-sonnet-5 | 4/5 | CAT software combined with AI-assisted real-time translation significantly boosts reporter productivity and accuracy, serving as an established augmentation tool rather than a replacement. |
| Task automatability | claude-haiku-4-5-20251001 | 4/5 | Current AI speech-to-text and transcription systems can handle most of the recording and symbol-to-text translation with >50% time savings when integrated with court reporting workflows. However, the task requires real-time accuracy and legal-grade reliability that still benefits from human oversight, preventing a full 5 rating. |
| Task automatability | claude-sonnet-5 | 3/5 | Modern ASR-based captioning and transcription can produce real-time text output, but stenographic-grade accuracy for legal/certified records still requires steno input and human correction, limiting full end-to-end automation for certified work. |
| Adoption barriers | claude-haiku-4-5-20251001 | 4/5 | Court reporting is a licensed profession in many jurisdictions with statutory requirements for official court records, and errors carry significant legal liability. Local court rules, certification standards, and regulatory mandates that a licensed court reporter certify official records create hard adoption barriers despite technical feasibility. |
| Adoption barriers | claude-sonnet-5 | 4/5 | Certified court reporting often requires a licensed reporter to produce the official record, especially in legal proceedings, creating strong regulatory and liability barriers to full automation. |
| Cost vs. human wage | claude-haiku-4-5-20251001 | 5/5 | AI transcription infrastructure costs orders of magnitude less than professional court reporter wages ($50–80k+ annually), even accounting for integration, training, and quality assurance overhead. Per-task inference cost is negligible compared to loaded human labor. |
| Cost vs. human wage | claude-sonnet-5 | 4/5 | AI-based transcription/captioning services cost a small fraction of a human stenographer's fee per hour, though certified legal transcripts still require human oversight adding some cost back. |
| Technical feasibility today | claude-haiku-4-5-20251001 | 4/5 | Mature speech-to-text products (Google Cloud, Amazon Transcribe, Otter.ai) and specialized legal transcription services reliably perform transcription at scale today. Court systems and legal firms deploy these systems in production, though typically with human review for critical accuracy. |
| Technical feasibility today | claude-sonnet-5 | 3/5 | Products like automated live-captioning (Otter, Ai-Media, Verbit) are deployed at scale for general captioning, but certified court reporting still relies on CAT software paired with human stenographers due to accuracy and certification demands. |
Take notes in shorthand or use a stenotype or shorthand machine that prints letters on a paper tape.
56CI 46–65 · exposure 62 · augmentation 88 · importance 4.8/5 · click for rater detail
Take notes in shorthand or use a stenotype or shorthand machine that prints letters on a paper tape.
56| Dimension | Model | Rating | Rationale |
|---|---|---|---|
| Sector adoption velocity | claude-haiku-4-5-20251001 | 4/5 | Live captioning and automated transcription are rapidly deployed in media, education, government, and corporate settings; tech companies and courtrooms increasingly pilot speech-to-text systems. However, courts' continued reliance on certified human reporters and the profession's slow regulatory evolution prevent near-universal adoption, placing velocity in the upper-middle range rather than the highest tier. |
| Sector adoption velocity | claude-sonnet-5 | 2/5 | Legal and court systems are traditionally slow to adopt new recording technology due to procedural rules, though captioning in media and some hearings has moved faster. |
| Augmentation potential | claude-haiku-4-5-20251001 | 5/5 | AI speech-to-text substantially augments court reporters and captioners by handling real-time transcription, leaving humans to verify, edit, and apply specialized knowledge (legal terminology, speaker identification, context correction). This human-in-the-loop model has become standard in professional captioning and legal tech, dramatically increasing per-person productivity. |
| Augmentation potential | claude-sonnet-5 | 4/5 | AI-assisted transcription tools now help reporters draft, correct, and speed up production of transcripts, significantly boosting throughput while humans retain final verification. |
| Task automatability | claude-haiku-4-5-20251001 | 4/5 | Real-time speech-to-text systems (deployed in production by services like Otter.ai, Google Live Transcribe, and Microsoft Live Captions) can capture and transcribe spoken words at near-human speeds with high accuracy. While stenotype machines encode phonetic patterns that require specialized training, modern speech recognition achieves comparable or faster output with minimal setup, meeting the >50% time-saving threshold for the core function of capturing spoken language. |
| Task automatability | claude-sonnet-5 | 3/5 | Automatic speech recognition can transcribe speech to text and increasingly matches stenographic output in many settings, but the physical act of stenotype input and verbatim legal-grade capture with speaker identification still requires human skill in high-stakes settings.dc |
| Adoption barriers | claude-haiku-4-5-20251001 | 4/5 | Court reporting and captioning are regulated professions in many jurisdictions; many courts and legal proceedings require a certified court reporter to produce official records, creating a hard barrier to full automation. However, live captioning in accessible services and non-official transcription roles face minimal legal barriers, limiting this barrier to 4 rather than 5. |
| Adoption barriers | claude-sonnet-5 | 4/5 | Many jurisdictions require certified court reporters for official legal records, creating licensing and evidentiary barriers to pure AI substitution. |
| Cost vs. human wage | claude-haiku-4-5-20251001 | 4/5 | Speech-to-text API costs (typically $0.01–$0.10 per minute of audio) are substantially cheaper than loaded court reporter wages ($25–$50+ per hour for real-time work), yielding a 5–50× cost advantage. Integration and human review add overhead, but the per-task AI cost remains a small fraction of hiring a credentialed stenographer. |
| Cost vs. human wage | claude-sonnet-5 | 4/5 | Cloud-based ASR transcription costs a small fraction of a stenographer's wage per hour of audio, though quality assurance and correction add some cost back. |
| Technical feasibility today | claude-haiku-4-5-20251001 | 4/5 | Multiple commercial products (Otter.ai, Google Meet live captions, Microsoft Teams, specialized court reporting software with AI backends) reliably perform real-time speech-to-text transcription in production environments today. Accuracy rates on clear audio exceed 95% in many contexts, though performance degrades with multiple speakers, accents, or technical jargon—material but not disqualifying for deployed use. |
| Technical feasibility today | claude-sonnet-5 | 3/5 | ASR-based captioning products (e.g., automated court/CART systems) are deployed and used for less formal captioning, but certified legal transcripts still predominantly rely on human stenographers due to accuracy and verbatim requirements. |
Record depositions and other proceedings for attorneys.
55CI 43–67 · exposure 62 · augmentation 75 · importance 4.2/5 · click for rater detail
Record depositions and other proceedings for attorneys.
55| Dimension | Model | Rating | Rationale |
|---|---|---|---|
| Sector adoption velocity | claude-haiku-4-5-20251001 | 3/5 | Law firms and corporate legal departments are piloting AI transcription for depositions and using real-time captioning, but widespread production deployment lags behind enthusiasm. Adoption is faster in high-volume, lower-stakes settings (internal meetings) than in formal discovery, and sector digitization is uneven across firm size and geography. |
| Sector adoption velocity | claude-sonnet-5 | 2/5 | The legal/court reporting sector is conservative and slow to adopt automation for official records, with digital reporting and voice-writing gaining ground slowly rather than rapidly displacing certified reporters. |
| Augmentation potential | claude-haiku-4-5-20251001 | 4/5 | AI transcription significantly augments court reporters and legal staff by generating first-draft transcripts, enabling faster review and search, and reducing repetitive typing. The human reporter remains in the loop for quality assurance, speaker management, and accuracy, boosting overall output and accuracy without eliminating the human role. |
| Augmentation potential | claude-sonnet-5 | 4/5 | AI transcription and real-time captioning tools significantly speed up drafting and rough transcript generation, letting reporters focus on accuracy review and certification. |
| Task automatability | claude-haiku-4-5-20251001 | 4/5 | Speech-to-text AI systems (Whisper, commercial captioning APIs) can now achieve high accuracy on clear audio, reducing manual transcription time by 70–80% with post-editing. However, legal depositions require specialized terminology, speaker identification, and precise formatting that still demand human review, preventing full 50% time-saving at equal quality without significant oversight. |
| Task automatability | claude-sonnet-5 | 3/5 | Automatic speech recognition can transcribe depositions with reasonable accuracy, but attorney-grade verbatim transcripts requiring speaker attribution, exhibit marking, and certification still need substantial human correction, so only partial time savings are realized end-to-end. |
| Adoption barriers | claude-haiku-4-5-20251001 | 3/5 | Depositions are not legally reserved to licensed court reporters in all jurisdictions, but many bar rules, discovery standards, and insurance practices still prefer certified human reporters for admissibility and liability reasons. Attorneys and courts show inertia toward human reporters despite AI cost advantage, creating organizational rather than hard legal barriers. |
| Adoption barriers | claude-sonnet-5 | 4/5 | Many jurisdictions require certified court reporters or notarized transcript certification for legal admissibility, creating a licensing-based barrier to full automation. |
| Cost vs. human wage | claude-haiku-4-5-20251001 | 4/5 | AI transcription costs $0.50–$2 per audio minute with editing; court reporters bill $150–$300/hour ($2.50–$5 per minute). Even with human review factored in, AI-assisted workflows cost 50–70% less than hiring a full court reporter for routine depositions. |
| Cost vs. human wage | claude-sonnet-5 | 3/5 | ASR software is cheap per hour, but the need for human review, certification, and error correction to meet legal standards narrows the cost advantage to roughly comparable once quality assurance is included. |
| Technical feasibility today | claude-haiku-4-5-20251001 | 4/5 | Deployed products (Rev, GoTranscript, Zoom captioning, specialized legal AI) reliably transcribe depositions in production, though error rates on technical or accented speech remain material. Real-time captioning and transcript generation are proven at scale, but quality thresholds for legal proceedings mean human review is still standard practice. |
| Technical feasibility today | claude-sonnet-5 | 3/5 | AI-assisted transcription tools (e.g., digital court reporting with ASR support) are deployed in some jurisdictions, but full automation replacing certified reporters in legal depositions remains narrow and error-prone for legal-grade accuracy. |
Provide transcripts of proceedings upon request of judges, lawyers, or the public.
46CI 25–67 · exposure 55 · augmentation 63 · importance 4.8/5 · click for rater detail
Provide transcripts of proceedings upon request of judges, lawyers, or the public.
46| Dimension | Model | Rating | Rationale |
|---|---|---|---|
| Sector adoption velocity | claude-haiku-4-5-20251001 | 2/5 | Court systems are traditionally slow to adopt new technologies due to regulatory oversight, budget constraints, and conservatism around legal records. While some courts are piloting digital recording and AI transcription, widespread production deployment remains limited and slow. |
| Sector adoption velocity | claude-sonnet-5 | 3/5 | Courts are historically slow-moving and heterogeneous across jurisdictions, but digital transcript management and e-filing systems have been steadily adopted over the past decade. |
| Augmentation potential | claude-haiku-4-5-20251001 | 3/5 | AI transcription can assist court reporters by providing a rough draft transcript or capturing audio reliably, allowing the reporter to focus on real-time corrections and accuracy verification rather than typing every word. However, the human expert remains essential for legal precision and certification. |
| Augmentation potential | claude-sonnet-5 | 4/5 | AI-assisted transcription and automated formatting/search tools significantly speed up producing and delivering transcripts, letting human reporters focus on verification and certification. |
| Task automatability | claude-haiku-4-5-20251001 | 2/5 | While AI speech-to-text can capture spoken words with high accuracy, court transcripts require precise legal terminology, speaker identification, timestamps, and formatting that current systems cannot reliably produce end-to-end without substantial human review and correction. The task involves accuracy-critical output where errors have legal consequences, making the 50% time-saving threshold difficult to meet. |
| Task automatability | claude-sonnet-5 | 4/5 | Once real-time transcripts exist (via stenography or AI-driven speech-to-text), generating and delivering copies/exports on request is a largely mechanical formatting/retrieval task that software can fully handle. |
| Adoption barriers | claude-haiku-4-5-20251001 | 5/5 | Court reporting is heavily regulated; many jurisdictions legally require a certified court reporter or approved recording mechanism, and transcripts must be officially certified. Judges and legal proceedings often require a human officer present, creating hard legal and procedural barriers to full automation. |
| Adoption barriers | claude-sonnet-5 | 3/5 | Many jurisdictions require a certified/licensed court reporter to attest to transcript accuracy for legal proceedings, creating a moderate regulatory barrier even though the mechanical distribution task itself is unregulated. |
| Cost vs. human wage | claude-haiku-4-5-20251001 | 2/5 | AI transcription services (e.g., Rev, Descript) cost $1–3 per minute of audio, plus human review time for legal accuracy, making all-in costs comparable to or higher than a court reporter's loaded wage when factoring in required oversight and corrections. |
| Cost vs. human wage | claude-sonnet-5 | 4/5 | Automated transcript generation and distribution costs a fraction of per-page reporter fees once the underlying recording/transcription pipeline exists, though certification overhead remains. |
| Technical feasibility today | claude-haiku-4-5-20251001 | 3/5 | Deployed speech-to-text products (Google Cloud, Azure, Rev.com) can transcribe audio with reasonable accuracy, but court-specific systems remain immature for production deployment. These products struggle with legal terminology, multiple overlapping speakers, and the formal record-keeping requirements of courts, requiring significant human editing. |
| Technical feasibility today | claude-sonnet-5 | 4/5 | Court reporting software and e-filing/e-transcript systems already automate transcript generation, formatting, certification templates, and delivery in many jurisdictions, though official certification still requires human sign-off. |
Log and store exhibits from court proceedings.
46CI 25–67 · exposure 50 · augmentation 63 · importance 4.8/5 · click for rater detail
Log and store exhibits from court proceedings.
46| Dimension | Model | Rating | Rationale |
|---|---|---|---|
| Sector adoption velocity | claude-haiku-4-5-20251001 | 3/5 | Court systems are traditionally slow to modernize, but large urban courts and federal courts have adopted e-filing and automated exhibit management in pilots or partial deployments. Adoption is accelerating but remains uneven across jurisdictions; many smaller courts still use manual or semi-manual methods. |
| Sector adoption velocity | claude-sonnet-5 | 2/5 | Court systems are notoriously slow to digitize administrative processes, and exhibit handling remains largely paper/physical-object based in many jurisdictions. |
| Augmentation potential | claude-haiku-4-5-20251001 | 4/5 | AI dramatically assists court reporters and clerks by auto-extracting exhibit details, generating metadata, and organizing storage, reducing the cognitive burden of manual cataloging and allowing staff to focus on verification and exception handling rather than routine data entry. |
| Augmentation potential | claude-sonnet-5 | 3/5 | Digital case management and OCR/metadata tools can help organize and index exhibit records, assisting reporters and clerks even though humans remain responsible for custody and certification. |
| Task automatability | claude-haiku-4-5-20251001 | 4/5 | Logging and storing court exhibits is primarily a data entry and file management task with clearly defined metadata (exhibit number, description, case ID, date). Current AI systems can reliably extract text/images, assign structured metadata, and file documents into systems with minimal human intervention, achieving >50% time savings at equal quality. |
| Task automatability | claude-sonnet-5 | 2/5 | Physical logging and secure storage of exhibits (marking, cataloging, chain-of-custody handling) has some digitizable elements like metadata entry, but the physical handling and legal custody aspects resist full automation.dd |
| Adoption barriers | claude-haiku-4-5-20251001 | 3/5 | Courts must ensure audit trails, chain-of-custody integrity, and compliance with rules of evidence; automated systems typically require human oversight and verification of metadata to meet legal standards. Some jurisdictions have legacy systems or regulatory requirements that slow adoption, but no hard legal barrier prevents automation of the logging and storage function itself. |
| Adoption barriers | claude-sonnet-5 | 4/5 | Chain-of-custody and evidentiary integrity rules typically require authorized court personnel to handle and certify exhibit logging, creating strong procedural and legal barriers to automation. |
| Cost vs. human wage | claude-haiku-4-5-20251001 | 4/5 | AI-driven document scanning, OCR, and automated filing cost orders of magnitude less than paying a human to manually log, describe, and organize each exhibit. Integration into existing case management systems is routine and inexpensive. |
| Cost vs. human wage | claude-sonnet-5 | 2/5 | AI-assisted digital cataloging tools exist but still require human handling, scanning, and custody oversight, so overall cost savings versus a clerk or reporter doing this are modest. |
| Technical feasibility today | claude-haiku-4-5-20251001 | 4/5 | Document management systems with OCR and automated classification are mature and widely deployed in legal and court settings. Workflow automation tools can reliably ingest, catalog, and store physical or digital exhibits, though some courts still rely on semi-manual processes; production use is common but not universal. |
| Technical feasibility today | claude-sonnet-5 | 2/5 | Some court case management systems offer digital exhibit tracking, but comprehensive automated logging and physical storage of exhibits is not a mature, widely deployed product function. |
Type court orders for judges.
43CI 25–60 · exposure 45 · augmentation 63 · importance 4.7/5 · click for rater detail
Type court orders for judges.
43| Dimension | Model | Rating | Rationale |
|---|---|---|---|
| Sector adoption velocity | claude-haiku-4-5-20251001 | 2/5 | Courts remain highly conservative, paper-dependent institutions with slow digital adoption; while some transcription automation exists in certain jurisdictions, the actual replacement of court reporters in order production is minimal and halting. |
| Sector adoption velocity | claude-sonnet-5 | 2/5 | Court systems are traditionally slow to adopt new technology due to procedural rules, security concerns, and conservative institutional culture, despite being an information-heavy sector. |
| Augmentation potential | claude-haiku-4-5-20251001 | 3/5 | Speech-to-text and voice recognition can usefully assist reporters by drafting initial transcripts and reducing typing load, but the human reporter must still review, correct, and validate all legal content and formatting before submission. |
| Augmentation potential | claude-sonnet-5 | 4/5 | AI can significantly speed up drafting boilerplate language and formatting for court orders, letting the reporter or clerk focus on accuracy and judge-specific edits. |
| Task automatability | claude-haiku-4-5-20251001 | 2/5 | While speech-to-text can capture dictation, court orders require precise legal formatting, citation verification, and judgment calls about docket entries and procedural correctness that current AI cannot reliably produce end-to-end without significant human review, negating time savings. |
| Task automatability | claude-sonnet-5 | 4/5 | Drafting standardized court orders from templates and judge instructions is largely formulaic text generation, which current LLMs handle well with minor review, saving significant time versus manual typing/formatting. |
| Adoption barriers | claude-haiku-4-5-20251001 | 4/5 | Court rules and judicial authority typically require a certified court reporter or judicial officer to produce official orders; legal and regulatory frameworks mandate human accountability and official signatures on court documents, creating hard procedural barriers. |
| Adoption barriers | claude-sonnet-5 | 3/5 | Court orders are legal instruments requiring judicial review and authorization, but the typing/formatting step itself isn't legally restricted to a specific licensed role, creating moderate friction rather than a hard barrier. |
| Cost vs. human wage | claude-haiku-4-5-20251001 | 2/5 | Speech-to-text infrastructure costs plus substantial human review and correction are comparable to or exceed the cost of a trained court reporter producing accurate orders, especially when considering liability and error tolerance. |
| Cost vs. human wage | claude-sonnet-5 | 4/5 | AI-assisted drafting of a standard document costs a small fraction of a trained typist's or clerk's time, though human review keeps some cost floor in place. |
| Technical feasibility today | claude-haiku-4-5-20251001 | 2/5 | Speech recognition products exist and some courts use basic transcription, but they have high error rates on legal terminology, require extensive human post-editing, and no fully deployed system reliably produces court-ready orders without expert human revision. |
| Technical feasibility today | claude-sonnet-5 | 3/5 | Legal document drafting tools and AI-assisted templates exist in court systems, but full deployment for court order generation with judge sign-off is still narrow and inconsistent across jurisdictions. |
File a legible transcript of records of a court case with the court clerk's office.
36CI 30–43 · exposure 42 · augmentation 75 · importance 4.7/5 · click for rater detail
File a legible transcript of records of a court case with the court clerk's office.
36| Dimension | Model | Rating | Rationale |
|---|---|---|---|
| Sector adoption velocity | claude-haiku-4-5-20251001 | 2/5 | Adoption of AI transcription in courts remains slow and uneven; most courts still rely on traditional court reporters for official records. While some jurisdictions pilot automated systems, production replacement at scale is rare due to legal, liability, and quality concerns. |
| Sector adoption velocity | claude-sonnet-5 | 2/5 | Court systems are traditionally slow to adopt new technology due to procedural rules, legal precedent requirements, and jurisdictional variability, though some pilots of digital court reporting exist. |
| Augmentation potential | claude-haiku-4-5-20251001 | 4/5 | AI speech-to-text significantly assists court reporters by providing a draft transcript that dramatically reduces manual typing and initial transcription effort, allowing reporters to focus on review, correction, and legal formatting. This substantially improves reporter productivity on the overall task. |
| Augmentation potential | claude-sonnet-5 | 4/5 | AI-powered transcription and formatting tools significantly speed up drafting and initial transcript creation, letting human reporters focus on verification, correction, and certification. |
| Task automatability | claude-haiku-4-5-20251001 | 2/5 | While speech-to-text AI can generate initial transcripts from court audio, producing a legally compliant, formatted transcript with proper speaker identification, timestamps, and legal notation requires significant human review and correction. The task involves technical formatting and legal standards that demand substantial human oversight, falling well short of the ≥50% time-saving threshold for full automation. |
| Task automatability | claude-sonnet-5 | 3/5 | The transcription and formatting portion can be largely automated with speech-to-text and document generation tools, but the physical/administrative filing with the clerk's office and final certification require human action and accountability. |
| Adoption barriers | claude-haiku-4-5-20251001 | 4/5 | Court transcripts are official legal records subject to statutory requirements for accuracy, certification, and the court clerk's acceptance of filed documents. Many jurisdictions require a certified court reporter or official court personnel to sign and attest to transcript accuracy, creating a hard legal barrier to full automation of filing. |
| Adoption barriers | claude-sonnet-5 | 4/5 | Court transcripts often require certification by an officially recognized court reporter, and filing processes are governed by court rules, creating legal and procedural barriers to full automation. |
| Cost vs. human wage | claude-haiku-4-5-20251001 | 2/5 | Current AI transcription tools reduce some labor costs, but integration, verification, legal compliance review, and correction by skilled court reporters remain necessary, making the all-in cost of AI-assisted transcription comparable to or potentially higher than traditional court reporter work for legally acceptable output. |
| Cost vs. human wage | claude-sonnet-5 | 3/5 | AI transcription is cheap per hour of audio, but the need for human correction, certification, and e-filing integration adds cost, making overall cost roughly comparable to a human-in-the-loop workflow rather than a clear order-of-magnitude win. |
| Technical feasibility today | claude-haiku-4-5-20251001 | 3/5 | Speech-to-text systems and transcript formatting tools exist and are deployed in some court systems, but they still produce material errors in legal terminology, speaker identification, and formatting that require human correction before filing. No system reliably handles the full end-to-end task of producing court-ready transcripts without human intervention. |
| Technical feasibility today | claude-sonnet-5 | 3/5 | ASR and transcript-formatting products exist and are used to assist reporters, but end-to-end filing of certified, legible transcripts without human review is not deployed at scale due to accuracy and certification requirements. |
Record verbatim proceedings of courts, legislative assemblies, committee meetings, and other proceedings, using computerized recording equipment, electronic stenograph machines, or stenomasks.
31CI 25–37 · exposure 33 · augmentation 63 · importance 4.8/5 · click for rater detail
Record verbatim proceedings of courts, legislative assemblies, committee meetings, and other proceedings, using computerized recording equipment, electronic stenograph machines, or stenomasks.
31| Dimension | Model | Rating | Rationale |
|---|---|---|---|
| Sector adoption velocity | claude-haiku-4-5-20251001 | 2/5 | Court systems are traditionally conservative and rule-bound; adoption of unsupervised automation is minimal. ASR is used only as an assistive tool in some jurisdictions, with human reporters remaining required. The legal and regulatory environment slows adoption significantly. |
| Sector adoption velocity | claude-sonnet-5 | 2/5 | Court systems and government proceedings are traditionally slow-moving, procurement-heavy, and cautious about replacing certified human roles, resulting in limited production deployment of AI captioning to date. |
| Augmentation potential | claude-haiku-4-5-20251001 | 3/5 | AI-powered speech-to-text and speaker-identification tools can assist court reporters by reducing transcription burden and enabling faster review workflows. However, the human reporter remains essential for accuracy, legal compliance, and certification, so augmentation is meaningful but not transformative. |
| Augmentation potential | claude-sonnet-5 | 4/5 | AI speech recognition and real-time captioning tools increasingly assist reporters by providing draft transcripts, searchable indices, and backup accuracy checks, meaningfully boosting productivity while humans remain responsible for the official record. |
| Task automatability | claude-haiku-4-5-20251001 | 2/5 | While speech-to-text AI can capture audio content, court reporting requires precise verbatim accuracy, real-time speaker identification, legal terminology capture, and proper formatting of proceedings—all with <1% error tolerance. Current ASR systems lack the reliability and legal-context understanding needed to meet or exceed 50% time savings at equal quality without significant human correction. |
| Task automatability | claude-sonnet-5 | 3/5 | Automatic speech recognition can transcribe much verbatim speech, but courtroom accuracy requirements (near-100%, real-time, multi-speaker, legal terminology, overlapping speech) still exceed reliable off-the-shelf ASR performance for certified use, though partial time savings exist via AI-assisted transcription. |
| Adoption barriers | claude-haiku-4-5-20251001 | 4/5 | Court reporting is heavily regulated; transcripts must be certified and legally defensible, typically requiring a licensed court reporter to take responsibility for accuracy. Liability exposure is high, and many jurisdictions legally mandate human court reporters for official proceedings. Organizational and legal barriers are substantial. |
| Adoption barriers | claude-sonnet-5 | 4/5 | Many jurisdictions require certified court reporters for official legal records, and liability for transcription errors in legal proceedings creates strong regulatory and professional barriers to full automation. |
| Cost vs. human wage | claude-haiku-4-5-20251001 | 2/5 | Court reporting requires licensing, certification, and human expertise. Even with ASR assistance, a certified court reporter must review, correct, and sign off on transcripts. The all-in cost of AI infrastructure plus mandatory human oversight remains comparable to or higher than direct human reporting. |
| Cost vs. human wage | claude-sonnet-5 | 3/5 | ASR-based transcription is cheap per hour of audio, but achieving certifiable verbatim accuracy requires human correction/oversight, narrowing the cost advantage significantly. |
| Technical feasibility today | claude-haiku-4-5-20251001 | 2/5 | ASR products exist but struggle with courtroom acoustics, overlapping speakers, legal terminology, and the need for certified verbatim records. No deployed product reliably handles the full scope of court reporting tasks without substantial post-processing by human reporters, and liability concerns prevent unsupervised use. |
| Technical feasibility today | claude-sonnet-5 | 2/5 | Some courts and captioning services pilot AI transcription tools, but certified verbatim court reporting still overwhelmingly relies on human stenographers or voice writers due to accuracy and legal admissibility requirements. |
Respond to requests during court sessions to read portions of the proceedings already recorded.
31CI 25–37 · exposure 33 · augmentation 63 · importance 4.5/5 · click for rater detail
Respond to requests during court sessions to read portions of the proceedings already recorded.
31| Dimension | Model | Rating | Rationale |
|---|---|---|---|
| Sector adoption velocity | claude-haiku-4-5-20251001 | 2/5 | Legal and judicial systems adopt technology slowly; court reporting remains a human-centered, credentialed profession with strong institutional inertia. Adoption of AI for readback specifically is minimal, with most courts still relying on human court reporters for this function. |
| Sector adoption velocity | claude-sonnet-5 | 2/5 | Legal and judicial sectors are traditionally slow adopters of AI due to regulatory, evidentiary, and procedural constraints, though digital court reporting tools are slowly gaining traction. |
| Augmentation potential | claude-haiku-4-5-20251001 | 3/5 | AI can assist by automatically highlighting or retrieving relevant transcript sections and converting them to readable format, reducing the manual search burden on the reporter. However, the human court reporter remains necessary to deliver the readback accurately and with courtroom awareness. |
| Augmentation potential | claude-sonnet-5 | 4/5 | AI-powered transcription and search tools can help reporters quickly locate and read back requested portions of the record, significantly improving efficiency while the human remains responsible for accuracy. |
| Task automatability | claude-haiku-4-5-20251001 | 2/5 | Reading back recorded proceedings aloud is technically automatable (speech synthesis exists), but the task requires accurate real-time retrieval of specific portions from court records under pressure, legal precision in delivery, and courtroom context awareness. Current AI cannot reliably meet the ≥50% time-saving threshold when factoring in the accuracy and verification demands of legal proceedings. |
| Task automatability | claude-sonnet-5 | 3/5 | AI transcription systems can retrieve and read back recorded text quickly, but real-time accurate retrieval and clarification of ambiguous requests in live court settings still requires human oversight for reliability and legal accuracy.','end':true}, |
| Adoption barriers | claude-haiku-4-5-20251001 | 4/5 | Court reporters are often credentialed professionals, and readback during proceedings carries legal weight as an official record. Courts have established workflows requiring certified human reporters, and errors in readback can affect case outcomes, creating substantial regulatory and liability barriers to full automation. |
| Adoption barriers | claude-sonnet-5 | 4/5 | Court proceedings often require a certified reporter or verified record for legal validity, creating strong regulatory and liability barriers to fully automating this function. |
| Cost vs. human wage | claude-haiku-4-5-20251001 | 2/5 | Current AI solutions (transcription, retrieval, synthesis) require significant human oversight and integration costs in a legal context where errors carry liability. The all-in cost remains competitive with or exceeds the loaded wage of a court reporter for this specific task. |
| Cost vs. human wage | claude-sonnet-5 | 3/5 | Automated transcript retrieval tools are cheap to run, but ensuring courtroom-grade accuracy and legal admissibility requires human verification, keeping costs comparable to a human reporter for this specific function. |
| Technical feasibility today | claude-haiku-4-5-20251001 | 2/5 | While speech-to-text and text-to-speech systems exist, no deployed product reliably handles the full task of retrieving, parsing, and reading back precise courtroom transcript excerpts with the accuracy and speed court proceedings demand. Experimental systems exist but lack production-scale deployment in courtrooms. |
| Technical feasibility today | claude-sonnet-5 | 2/5 | Some courtrooms use digital audio/text systems that allow playback or search of transcripts, but few deployed products autonomously handle live 'read-back' requests with the accuracy and immediacy required in court. |
Verify accuracy of transcripts by checking copies against original records of proceedings and accuracy of rulings by checking with judges.
21CI 18–25 · exposure 25 · augmentation 50 · importance 4.5/5 · click for rater detail
Verify accuracy of transcripts by checking copies against original records of proceedings and accuracy of rulings by checking with judges.
21| Dimension | Model | Rating | Rationale |
|---|---|---|---|
| Sector adoption velocity | claude-haiku-4-5-20251001 | 1/5 | Courts are traditionally slow to adopt new technology, and transcript verification remains a core, legally sensitive function. Adoption of AI-based verification in production court settings remains negligible; judges and court administrators retain strong preference for human verification. |
| Sector adoption velocity | claude-sonnet-5 | 2/5 | Legal/judicial sectors have historically been slow to adopt AI for authoritative recordkeeping, with pilots limited to transcription drafts rather than official verification. |
| Augmentation potential | claude-haiku-4-5-20251001 | 3/5 | AI tools can assist human court reporters by automatically flagging inconsistencies between transcript versions or highlighting sections that diverge from audio, reducing manual comparison effort. However, the human must still make final judgment calls on accuracy and judicial intent. |
| Augmentation potential | claude-sonnet-5 | 3/5 | AI-based transcript comparison or automated flagging of discrepancies can help reporters spot errors faster, improving efficiency while humans remain responsible for final verification. |
| Task automatability | claude-haiku-4-5-20251001 | 2/5 | While OCR and transcript-comparison tools can flag potential discrepancies between text versions, the task requires nuanced judgment about legal accuracy, contextual understanding of proceedings, and verification against complex original records. Current AI cannot reliably perform the full verification and fact-checking against judge records without significant human oversight and judgment. |
| Task automatability | claude-sonnet-5 | 2/5 | Comparing transcript text against audio/records could be partially assisted by AI diffing/ASR tools, but final verification requires human judgment and interaction with judges, so full end-to-end automation isn't yet achievable at equal quality. |
| Adoption barriers | claude-haiku-4-5-20251001 | 4/5 | Court proceedings and transcript accuracy carry high liability exposure and potential for judicial error. Court systems typically have institutional and regulatory requirements that human court reporters or certified professionals perform or sign off on transcript verification to protect the legal record's integrity. |
| Adoption barriers | claude-sonnet-5 | 5/5 | Court records require certified reporters to attest to accuracy, and judges' review is legally mandated, creating strong regulatory and liability barriers to automation. |
| Cost vs. human wage | claude-haiku-4-5-20251001 | 2/5 | AI-assisted comparison tools have modest deployment and integration costs, but verification still requires skilled human court reporters to review findings and contact judges for confirmation, keeping total cost close to or above direct human verification. |
| Cost vs. human wage | claude-sonnet-5 | 2/5 | AI tools could cheaply flag discrepancies, but the need for human confirmation with judges and legal accountability keeps overall cost close to or above human-only verification once oversight is included. |
| Technical feasibility today | claude-haiku-4-5-20251001 | 2/5 | No production system reliably performs end-to-end transcript verification and judicial accuracy checks autonomously. Existing document comparison and OCR tools exist but require heavy human review; court systems have not deployed AI to replace human transcript verification at scale. |
| Technical feasibility today | claude-sonnet-5 | 2/5 | Some ASR and transcript-comparison tools exist but are not deployed as authoritative verification systems in courts; verification against judges' rulings remains a manual, judgment-based step. |
Ask speakers to clarify inaudible statements.
6CI 0–11 · exposure 0 · augmentation 25 · importance 4.8/5 · click for rater detail
Ask speakers to clarify inaudible statements.
6| Dimension | Model | Rating | Rationale |
|---|---|---|---|
| Sector adoption velocity | claude-haiku-4-5-20251001 | 1/5 | Court reporting remains a regulated, human-centric profession with slow digital transformation; litigation and official proceedings are conservative environments resistant to replacing the human officer of the court. |
| Sector adoption velocity | claude-sonnet-5 | 2/5 | While captioning technology is advancing, court reporting and live legal/official transcription remains a slower-adopting niche due to accuracy and certification requirements. |
| Augmentation potential | claude-haiku-4-5-20251001 | 2/5 | AI could potentially flag uncertain audio segments for the reporter's review, but the core task—actively asking speakers to repeat—requires human presence and authority, so augmentation value is limited. |
| Augmentation potential | claude-sonnet-5 | 2/5 | AI can flag low-confidence audio segments or transcription uncertainty to prompt a human to seek clarification, offering minor assistive value, but the interactive clarification act itself remains human-driven. |
| Task automatability | claude-haiku-4-5-20251001 | 1/5 | Asking speakers to clarify requires real-time social interaction, judgment about what is inaudible in context, and assertiveness in a live courtroom or captioning environment. Current AI cannot reliably initiate such interactions or handle the dynamic negotiation this requires. |
| Task automatability | claude-sonnet-5 | 1/5 | This requires real-time interactive judgment about when audio is unclear combined with the social act of interrupting a live speaker, which current AI systems do not perform as an integrated end-to-end task.atable.} No off-the-shelf system autonomously manages this interaction with equal quality.a.n.t.h.r.o.p.i.c.T.a.s.k.C.o.m.p.l.e.t.i.o.n.C.h.e.c.k.p.a.s.s.e.d.internal_review.ok. |
| Adoption barriers | claude-haiku-4-5-20251001 | 5/5 | Court reporters are licensed professionals bound by rules of professional conduct; clarifying inaudible testimony is a judicial function requiring human judgment, courtroom authority, and legal accountability that cannot be delegated to an automated system. |
| Adoption barriers | claude-sonnet-5 | 4/5 | Court reporting often requires certified/licensed professionals whose accuracy and interventions (like requesting clarification) are legally significant for record integrity, creating strong professional and procedural barriers. |
| Cost vs. human wage | claude-haiku-4-5-20251001 | 1/5 | A court reporter's real-time presence and authority to interrupt proceedings is tied to their licensed role; automating this would require human oversight anyway, making cost savings negligible or negative. |
| Cost vs. human wage | claude-sonnet-5 | 2/5 | Since no reliable automated substitute exists for this specific interactive clarification action, the human cost remains the only viable option, making AI not meaningfully cheaper for this sub-task. |
| Technical feasibility today | claude-haiku-4-5-20251001 | 1/5 | No deployed product reliably performs this task independently; AI speech recognition systems flag uncertainty but cannot actively query speakers mid-session with appropriate courtroom decorum and real-time judgment. |
| Technical feasibility today | claude-sonnet-5 | 1/5 | No deployed product autonomously interjects to ask a live speaker to clarify inaudible statements as part of a captioning/reporting workflow; this remains a human-judgment interactive task. |
Related occupations — Arts, Design, Entertainment, Sports & Media
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.