Interpreters and Translators
27-3091.00Median wage $60,170/yr52,060 employed (US)Rank #139 of 923 scored · top 15% by substitution
Interpret oral or sign language, or translate written text from one language into another.
Sub-scores
0–100 · band = confidence interval from rater disagreement
Substitution42
Exposure36
Augmentation71
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
17 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
12%
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%35
panel mean rating 2.4/5 → substitution pressure 35/100
Technical feasibility todayw 20%38
panel mean rating 2.5/5 → substitution pressure 38/100
Cost vs. human wagew 15%51
panel mean rating 3.0/5 → substitution pressure 51/100
Adoption barriersw 20%inverted — strong barriers lower the score50
panel mean rating 3.0/5 (barrier strength) → substitution pressure 50/100
Sector adoption velocityw 10%42
panel mean rating 2.7/5 → substitution pressure 42/100
Task breakdown (17 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.
Refer to reference materials, such as dictionaries, lexicons, encyclopedias, and computerized terminology banks, as needed to ensure translation accuracy.
91CI 84–97 · exposure 92 · augmentation 100 · importance 4.4/5 · click for rater detail
Refer to reference materials, such as dictionaries, lexicons, encyclopedias, and computerized terminology banks, as needed to ensure translation accuracy.
91| Dimension | Model | Rating | Rationale |
|---|---|---|---|
| Sector adoption velocity | claude-haiku-4-5-20251001 | 4/5 | Professional translation services, localization firms, and multilingual enterprises are actively adopting AI-assisted and AI-driven translation with integrated terminology management. CAT tools with automated reference lookup are now standard in the professional translation industry. |
| Sector adoption velocity | claude-sonnet-5 | 4/5 | Translation is an information-sector task with fast, deep AI tool adoption industry-wide, including integrated terminology management in CAT tools. |
| Augmentation potential | claude-haiku-4-5-20251001 | 5/5 | AI-powered reference lookup and terminology suggestion significantly augments human translators by eliminating manual dictionary searches and flagging term inconsistencies in real time, substantially raising translator productivity while the human remains in control of quality. |
| Augmentation potential | claude-sonnet-5 | 5/5 | AI-powered terminology suggestion and glossary integration significantly speeds up and improves consistency for human translators who remain responsible for final quality. |
| Task automatability | claude-haiku-4-5-20251001 | 4/5 | Current AI systems can autonomously consult reference materials, databases, and terminology banks as part of end-to-end translation workflows, achieving significant time savings. The main constraint is that ensuring translation accuracy still requires human review in many contexts, but the reference-lookup component itself is nearly fully automatable. |
| Task automatability | claude-sonnet-5 | 5/5 | Modern LLMs and MT systems internalize terminology lookup and reference consultation as an integral part of translation, effectively automating this sub-task with equal or better speed than manual reference checking. |
| Adoption barriers | claude-haiku-4-5-20251001 | 2/5 | There are minimal legal or regulatory barriers to automating reference material consultation. The main friction is customer preference for human translation in high-stakes domains and organizational inertia, but no licensing requirement prevents automation of this specific task. |
| Adoption barriers | claude-sonnet-5 | 1/5 | This is a research/reference-support task with no licensing, liability, or human-contact requirement blocking AI use. |
| Cost vs. human wage | claude-haiku-4-5-20251001 | 5/5 | AI systems can perform reference consultation and term lookup at near-zero marginal cost per task, orders of magnitude cheaper than human translators who must manually search dictionaries and databases for specialized or ambiguous terms. |
| Cost vs. human wage | claude-sonnet-5 | 5/5 | Automated terminology lookup and integration into translation output costs fractions of a cent versus the time a human translator would spend manually consulting dictionaries and lexicons. |
| Technical feasibility today | claude-haiku-4-5-20251001 | 5/5 | Deployed translation products (Google Translate, DeepL, professional CAT tools) routinely integrate terminology databases and reference lookups automatically during translation. This is a standard, production-proven capability at scale across multiple platforms. |
| Technical feasibility today | claude-sonnet-5 | 5/5 | Production translation tools (DeepL, Google Translate, GPT-based CAT tool plugins) routinely incorporate glossaries, terminology banks, and translation memories reliably at scale. |
Read written materials, such as legal documents, scientific works, or news reports, and rewrite material into specified languages.
79CI 79–79 · exposure 75 · augmentation 88 · importance 3.7/5 · click for rater detail
Read written materials, such as legal documents, scientific works, or news reports, and rewrite material into specified languages.
79| Dimension | Model | Rating | Rationale |
|---|---|---|---|
| Sector adoption velocity | claude-haiku-4-5-20251001 | 4/5 | High adoption is evident across information, publishing, and global commerce sectors; translation automation is mainstream in digital workflows, though some regulated or high-stakes contexts (legal certification) remain slower to shift. |
| Sector adoption velocity | claude-sonnet-5 | 4/5 | Translation/localization industry has rapidly integrated MT and AI post-editing workflows, with major language service providers now using AI-first pipelines with human post-editors. |
| Augmentation potential | claude-haiku-4-5-20251001 | 4/5 | AI translation tools substantially augment human translators by handling drafting, terminology lookup, and batch processing, allowing humans to focus on review, cultural adaptation, and quality assurance rather than raw translation. |
| Augmentation potential | claude-sonnet-5 | 5/5 | AI-assisted translation (MT plus post-editing) is now the dominant augmented workflow in the industry, dramatically increasing translator throughput while retaining human review for accuracy and nuance. |
| Task automatability | claude-haiku-4-5-20251001 | 4/5 | Neural machine translation (NMT) systems can now produce fluent, contextually appropriate translations of written materials at speeds far exceeding human capability, achieving >50% time savings at comparable quality for many document types. However, specialized legal, scientific, or culturally nuanced content still requires human review, preventing a full 5 rating. |
| Task automatability | claude-sonnet-5 | 4/5 | Machine translation systems and LLMs can now translate most document types with high fluency and reasonable accuracy, achieving substantial time savings for drafting, though high-stakes legal/technical texts still need human review. |
| Adoption barriers | claude-haiku-4-5-20251001 | 2/5 | While some legal or certified translation contexts may require a licensed human's signature, most written material translation has weak barriers; organizations adopt MT freely without authorization or liability constraints that would block substitution. |
| Adoption barriers | claude-sonnet-5 | 2/5 | Some domains (certified legal/medical translations, sworn court documents) require certified human translators for liability and legal-validity reasons, but most general translation work has no licensing requirement. |
| Cost vs. human wage | claude-haiku-4-5-20251001 | 5/5 | Machine translation inference costs are orders of magnitude cheaper than professional human translation when accounting for speed and scale, even with oversight overhead included. |
| Cost vs. human wage | claude-sonnet-5 | 5/5 | AI translation costs fractions of a cent per word versus professional human translator rates of $0.10-$0.30/word, an order-of-magnitude-plus cost advantage for initial drafting. |
| Technical feasibility today | claude-haiku-4-5-20251001 | 4/5 | Production-grade translation services (DeepL, Google Translate, Microsoft Translator) are deployed at enterprise scale and handle diverse written materials reliably. Error rates remain material in specialized domains, and quality varies by language pair, keeping this below a 5. |
| Technical feasibility today | claude-sonnet-5 | 4/5 | Production translation tools (DeepL, Google Translate, GPT-4-based systems) are widely deployed in enterprises and translation agencies today for draft translation across many language pairs, though quality varies by domain and language rarity. |
Proofread, edit, and revise translated materials.
67CI 55–79 · exposure 62 · augmentation 88 · importance 3.9/5 · click for rater detail
Proofread, edit, and revise translated materials.
67| Dimension | Model | Rating | Rationale |
|---|---|---|---|
| Sector adoption velocity | claude-haiku-4-5-20251001 | 3/5 | Professional translation services and publishing firms increasingly integrate AI proofreading tools into workflows, but adoption remains primarily assistive (augmentation mode) rather than replacement; independent translators and enterprise language services show gradual uptake, not yet deep displacement. |
| Sector adoption velocity | claude-sonnet-5 | 4/5 | Language services and localization industries have rapidly integrated AI/MT post-editing workflows, with LLM-based tools now standard in many translation agencies and enterprise localization pipelines. |
| Augmentation potential | claude-haiku-4-5-20251001 | 4/5 | AI-powered editing tools substantially enhance translator productivity by flagging potential errors, suggesting terminology consistency, and accelerating first-pass quality checks, enabling human translators to focus on nuanced judgment while the AI surfaces routine issues at scale. |
| Augmentation potential | claude-sonnet-5 | 5/5 | AI tools dramatically speed up proofreading and revision by flagging errors, suggesting corrections, and ensuring consistency, while human translators retain final judgment and quality control. |
| Task automatability | claude-haiku-4-5-20251001 | 3/5 | AI language models can detect and correct many surface-level errors (grammar, spelling, terminology consistency) and some semantic issues in translated text, but require human judgment for cultural nuance, context-dependent accuracy, and subjective quality trade-offs. Full automation would require domain expertise and quality control oversight that pushes beyond the 50% time-saving bar for many specialized content types. |
| Task automatability | claude-sonnet-5 | 4/5 | AI translation and editing tools can now proofread and revise translated text for grammar, fluency, and terminology consistency with substantial time savings, though nuanced cultural or high-stakes legal/literary revision still benefits from human review. |
| Adoption barriers | claude-haiku-4-5-20251001 | 2/5 | Translation quality directly impacts client outcomes and legal/commercial liability, creating risk-averse organizational culture and customer preference for human review, but no hard legal licensing barrier mandates a human sign-off on the editing step itself. |
| Adoption barriers | claude-sonnet-5 | 2/5 | Most translation work has no licensing requirement, though certified/legal/medical translations may require human sign-off, creating moderate friction in specific niches. |
| Cost vs. human wage | claude-haiku-4-5-20251001 | 3/5 | AI inference and integration costs for proofreading are substantially lower than human labor, but oversight and human validation of AI suggestions often consume enough time that all-in cost becomes roughly comparable to hiring junior editors for initial review and quality checks. |
| Cost vs. human wage | claude-sonnet-5 | 5/5 | AI-assisted proofreading/editing costs a fraction of a cent to a few cents per document versus hourly human translator/editor rates, representing an order-of-magnitude cost advantage. |
| Technical feasibility today | claude-haiku-4-5-20251001 | 3/5 | Deployed products (grammar checkers, AI editing assistants, translation QA tools) exist and perform adequately on routine edits, but show material gaps in detecting subtle meaning shifts, register mismatches, and domain-specific errors. Most are used as assistants rather than autonomous end-to-end solutions in professional translation workflows. |
| Technical feasibility today | claude-sonnet-5 | 4/5 | Deployed products (DeepL, Google Translate post-editing tools, LLM-based CAT tool plugins) are widely used in production for translation QA and revision, though error rates remain nontrivial for specialized domains. |
Check translations of technical terms and terminology to ensure that they are accurate and remain consistent throughout translation revisions.
56CI 37–75 · exposure 55 · augmentation 88 · importance 4.3/5 · click for rater detail
Check translations of technical terms and terminology to ensure that they are accurate and remain consistent throughout translation revisions.
56| Dimension | Model | Rating | Rationale |
|---|---|---|---|
| Sector adoption velocity | claude-haiku-4-5-20251001 | 3/5 | Translation and localization firms have adopted CAT tools and AI-assisted QA at pilots and early-production stages, but widespread displacement remains limited because quality-critical work still demands human expertise and client trust in named individuals. |
| Sector adoption velocity | claude-sonnet-5 | 4/5 | Translation and localization industries have rapidly integrated AI-assisted QA and terminology tools, with widespread production use in language service providers. |
| Augmentation potential | claude-haiku-4-5-20251001 | 4/5 | AI excels at consistency-checking and terminology database flagging, significantly raising translator productivity by automating tedious scanning for variant usages and suggesting corrections while the human validates and makes final decisions. |
| Augmentation potential | claude-sonnet-5 | 5/5 | AI terminology checkers and consistency tools significantly boost translator productivity by flagging inconsistencies instantly while the human retains final judgment. |
| Task automatability | claude-haiku-4-5-20251001 | 2/5 | AI systems can flag inconsistencies in terminology use and suggest corrections with reasonable accuracy, but they struggle with domain-specific technical context, rare terminology, and nuanced equivalency across languages—tasks requiring deep expertise that humans still must verify. |
| Task automatability | claude-sonnet-5 | 4/5 | Terminology consistency checking against glossaries/translation memories is a pattern-matching task that current LLM and CAT-tool integrations handle well, saving significant time versus manual checking. |
| Adoption barriers | claude-haiku-4-5-20251001 | 3/5 | Professional translation often involves client contracts specifying human review and sign-off, and liability for mistranslations creates organizational preference for human accountability; however, no strict legal licensing barrier exists to full automation. |
| Adoption barriers | claude-sonnet-5 | 2/5 | No licensing requirement for this specific sub-task, though quality/liability concerns in specialized domains (legal, medical) create some organizational caution before fully trusting automated checks. |
| Cost vs. human wage | claude-haiku-4-5-20251001 | 2/5 | AI-assisted terminology checking tools are cheap per document, but full end-to-end replacement requires human oversight due to error rates, making combined costs competitive with rather than substantially cheaper than skilled human checkers. |
| Cost vs. human wage | claude-sonnet-5 | 4/5 | Automated terminology QA checks run at a fraction of the cost of manual line-by-line human verification, though some human oversight remains necessary. |
| Technical feasibility today | claude-haiku-4-5-20251001 | 3/5 | AI translation quality-assurance tools and terminology databases exist (e.g., CAT tools with AI-assisted consistency checking, neural MT with term-lock features), but they produce material false positives/negatives and require human review, limiting autonomous reliability in production. |
| Technical feasibility today | claude-sonnet-5 | 4/5 | CAT tools (SDL Trados, memoQ) with QA modules and AI-assisted terminology checkers are widely deployed in production translation workflows today, though edge cases still require human review. |
Compile terminology and information to be used in translations, including technical terms such as those for legal or medical material.
52CI 25–79 · exposure 50 · augmentation 88 · importance 4.5/5 · click for rater detail
Compile terminology and information to be used in translations, including technical terms such as those for legal or medical material.
52| Dimension | Model | Rating | Rationale |
|---|---|---|---|
| Sector adoption velocity | claude-haiku-4-5-20251001 | 2/5 | Translation agencies and language service providers do use terminology management and CAT tools, but adoption of autonomous AI-driven compilation remains cautious due to accuracy and liability concerns, particularly in regulated sectors. Most adoption is assistive, not replacement. |
| Sector adoption velocity | claude-sonnet-5 | 4/5 | Language services and professional translation is a digitized, fast-adopting sector where CAT tools and AI-assisted terminology management are now standard practice. |
| Augmentation potential | claude-haiku-4-5-20251001 | 4/5 | AI excels at rapidly surfacing candidate terminology, building draft glossaries, detecting inconsistencies, and flagging missing terms. Translators using AI-augmented tools can compile richer, more consistent terminology faster, substantially raising productivity while retaining expert judgment on validation. |
| Augmentation potential | claude-sonnet-5 | 5/5 | AI dramatically speeds up terminology research and glossary building, letting translators focus on nuanced usage and context while AI handles bulk compilation. |
| Task automatability | claude-haiku-4-5-20251001 | 2/5 | Terminology compilation has algorithmic components (extracting terms, identifying glossaries, cross-referencing), but requires domain expertise to validate accuracy, assess context-specificity, and curate quality. Current AI can draft terminology lists but human review of legal/medical terms is essential, falling short of the 50% time-saving threshold for equal quality. |
| Task automatability | claude-sonnet-5 | 4/5 | AI systems can rapidly compile, cross-reference, and organize domain-specific terminology (legal/medical glossaries, term banks) with high speed, though final validation for specialized accuracy often still benefits from expert review. |
| Adoption barriers | claude-haiku-4-5-20251001 | 4/5 | Medical and legal translation face regulatory oversight, liability for errors, and professional credentialing requirements. Terminological accuracy directly impacts legal/clinical outcomes, creating strong organizational and legal friction against full automation without human expert sign-off. |
| Adoption barriers | claude-sonnet-5 | 2/5 | No licensing requirement for terminology compilation itself, though legal/medical domains carry some liability concern if inaccurate terms propagate into final translations. |
| Cost vs. human wage | claude-haiku-4-5-20251001 | 2/5 | AI-assisted terminology compilation (CAT tools, extraction engines) can reduce labor, but domain experts must still validate results, and the overhead of integration and quality-checking keeps total costs near or slightly below equivalent human effort for high-stakes translations. |
| Cost vs. human wage | claude-sonnet-5 | 5/5 | Automated term extraction and glossary compilation from corpora is vastly cheaper than manual terminology research by a human translator, especially at scale. |
| Technical feasibility today | claude-haiku-4-5-20251001 | 2/5 | Products exist (CAT tools, terminology databases, some NMT systems with custom glossaries) but they lack reliable end-to-end capability for compiling authoritative technical terminology, especially in regulated domains. Manual curation and validation remain required, limiting production reliability. |
| Technical feasibility today | claude-sonnet-5 | 4/5 | Terminology management tools with AI-assisted extraction (e.g., SDL Trados, memoQ, DeepL glossary features) are already deployed in production translation workflows for building term bases. |
Translate messages simultaneously or consecutively into specified languages, orally or by using hand signs, maintaining message content, context, and style as much as possible.
51CI 40–61 · exposure 38 · augmentation 75 · importance 4.7/5 · click for rater detail
Translate messages simultaneously or consecutively into specified languages, orally or by using hand signs, maintaining message content, context, and style as much as possible.
51| Dimension | Model | Rating | Rationale |
|---|---|---|---|
| Sector adoption velocity | claude-haiku-4-5-20251001 | 4/5 | Machine translation adoption is rapid and pervasive across digital, information, and service sectors. Real-time translation features ship in mainstream products (Google Translate, Microsoft Teams, Zoom); casual and business use are ubiquitous. However, high-stakes professional interpretation (conference, medical, legal) lags behind, preventing a full 5 rating. |
| Sector adoption velocity | claude-sonnet-5 | 3/5 | Adoption is growing quickly in business and travel contexts but remains slow in high-stakes professional interpretation (legal, medical, diplomatic) where human interpreters are still standard practice. |
| Augmentation potential | claude-haiku-4-5-20251001 | 4/5 | AI translation tools significantly assist human interpreters and translators by providing rapid drafts, terminology databases, and real-time suggestions. Interpreters use machine-generated text to accelerate preparation and reduce cognitive load, raising productivity while the human remains essential for quality control and nuance. |
| Augmentation potential | claude-sonnet-5 | 4/5 | AI tools are widely used to assist human interpreters with terminology look-up, real-time subtitling, and draft translations, meaningfully boosting speed and consistency while the human retains final judgment. |
| Task automatability | claude-haiku-4-5-20251001 | 2/5 | While machine translation has improved dramatically, simultaneous/consecutive interpretation requires real-time processing, maintaining nuance, context, and style—tasks where current AI still produces frequent errors and tone/cultural mismatches. End-to-end replacement meeting the 50% time-saving bar at equal quality remains unachieved, especially for specialized or culturally sensitive content. |
| Task automatability | claude-sonnet-5 | 3/5 | AI speech-to-speech translation tools can handle simultaneous/consecutive translation for common language pairs and straightforward content with significant time savings, but nuance, style preservation, cultural context, and sign language remain weak points, especially for high-stakes or specialized settings. |
| Adoption barriers | claude-haiku-4-5-20251001 | 2/5 | Interpretation of legal, medical, or diplomatic content often requires human certification and liability accountability; however, barriers are not absolute—many organizations and platforms accept machine translation for lower-stakes communication. Customer preference and regulatory requirements in some sectors (healthcare, courts) still favor licensed humans, but these are not hard legal blocks in all contexts. |
| Adoption barriers | claude-sonnet-5 | 3/5 | Court, medical, and diplomatic interpretation often require certified/licensed human interpreters and carry liability for mistranslation, creating meaningful friction, though casual or business interpretation faces fewer legal barriers. |
| Cost vs. human wage | claude-haiku-4-5-20251001 | 4/5 | Machine translation inference is extremely cheap (fractions of a cent per passage), vastly lower than hiring a human interpreter (often $50–$150+ per hour). Even accounting for integration and oversight, the cost ratio heavily favors AI, though post-editing by a human may still be required for high-stakes content. |
| Cost vs. human wage | claude-sonnet-5 | 4/5 | AI translation tools cost a small fraction of professional interpreter fees per unit of output, though oversight and correction for high-stakes contexts add cost, keeping it just short of the top tier. |
| Technical feasibility today | claude-haiku-4-5-20251001 | 2/5 | Commercial neural MT systems like Google Translate and DeepL handle static text adequately but struggle with live, bidirectional interpretation. Real-time speech-to-speech systems exist (e.g., Google Live Translate) but deploy in limited contexts with known error rates; no mature production system reliably matches human interpreter quality across language pairs and domains. |
| Technical feasibility today | claude-sonnet-5 | 3/5 | Products like Google Translate, Microsoft Translator, and conference interpretation AI tools are deployed and used in real settings, but they still show material error rates in idiomatic, emotionally nuanced, or technical speech, and sign language interpretation via AI is far less mature. |
Listen to speakers' statements to determine meanings and to prepare translations, using electronic listening systems as necessary.
49CI 45–54 · exposure 42 · augmentation 75 · importance 4.5/5 · click for rater detail
Listen to speakers' statements to determine meanings and to prepare translations, using electronic listening systems as necessary.
49| Dimension | Model | Rating | Rationale |
|---|---|---|---|
| Sector adoption velocity | claude-haiku-4-5-20251001 | 3/5 | AI-assisted translation is in broad but shallow adoption: many organizations pilot real-time captioning and machine translation, but professional interpreter displacement remains limited. Sectors like tech and customer service adopt fast; legal and healthcare lag significantly. |
| Sector adoption velocity | claude-sonnet-5 | 3/5 | Translation and interpretation services are adopting AI at a moderate pace—automated captioning and translation tools are common, but professional interpretation retains heavy human reliance in regulated and high-stakes domains. |
| Augmentation potential | claude-haiku-4-5-20251001 | 4/5 | AI speech-to-text and translation engine outputs serve as strong real-time aids to human interpreters and translators—reducing cognitive load, providing immediate reference versions, and enabling faster post-editing or real-time overlay. This is a mature augmentation pattern in production today. |
| Augmentation potential | claude-sonnet-5 | 4/5 | AI listening/transcription tools significantly speed up interpreters' draft preparation and real-time support, letting humans focus on refining nuance and accuracy rather than transcription from scratch. |
| Task automatability | claude-haiku-4-5-20251001 | 2/5 | Current AI (speech-to-text + neural machine translation) can handle segments of this task, but real-time interpretation requires nuanced understanding of context, idiomatic meaning, and cultural references that AI still struggles with reliably. End-to-end replacement would require <50% time savings at equal quality due to frequent human correction loops needed. |
| Task automatability | claude-sonnet-5 | 3/5 | AI speech-to-text and translation systems can handle straightforward spoken content reasonably well, but nuanced meaning, idiom, tone, and context in live interpretation still require human judgment for high-stakes or complex settings, limiting full end-to-end substitution to about half the workload. |
| Adoption barriers | claude-haiku-4-5-20251001 | 3/5 | While there are no strict licensing requirements preventing AI use, client preference for human interpreters in legal, medical, and diplomatic contexts, liability exposure for mistranslations, and organizational inertia provide moderate friction. High-stakes contexts (courtroom, medical) often still mandate human interpreters by custom or policy. |
| Adoption barriers | claude-sonnet-5 | 3/5 | Court, medical, and diplomatic interpretation often require certified human interpreters for legal validity and liability reasons, though many everyday or business contexts have no such requirement. |
| Cost vs. human wage | claude-haiku-4-5-20251001 | 4/5 | AI inference cost for speech-to-text and translation is now very cheap per utterance, and cloud-based systems scale efficiently. Even accounting for integration and oversight, the marginal cost per task is substantially below the loaded wage of a professional interpreter or translator. |
| Cost vs. human wage | claude-sonnet-5 | 4/5 | AI-based transcription and translation tools cost a small fraction of a professional interpreter's hourly rate, though oversight and correction for critical accuracy needs add some cost. |
| Technical feasibility today | claude-haiku-4-5-20251001 | 3/5 | Machine translation and speech-to-text products exist and are widely deployed (Google Translate, Microsoft, AWS), but their error rates on nuanced spoken language remain material, especially for specialized or real-time interpretation contexts. Production use is typically for informal communication, not high-stakes settings. |
| Technical feasibility today | claude-sonnet-5 | 3/5 | Deployed products (e.g., real-time translation earpieces, Google/Microsoft speech translation, Zoom AI interpretation) exist and are used in some settings, but accuracy for nuanced or specialized speech remains inconsistent, restricting reliable production use to lower-stakes contexts. |
Adapt software and accompanying technical documents to another language and culture.
49CI 37–61 · exposure 42 · augmentation 88 · importance 4.2/5 · click for rater detail
Adapt software and accompanying technical documents to another language and culture.
49| Dimension | Model | Rating | Rationale |
|---|---|---|---|
| Sector adoption velocity | claude-haiku-4-5-20251001 | 3/5 | Software and tech companies use AI translation tools extensively in pilot and production workflows, but human translators remain embedded in quality assurance and cultural review steps; adoption is widespread but not replacement-focused. |
| Sector adoption velocity | claude-sonnet-5 | 4/5 | Localization industry has rapidly adopted MT plus post-editing workflows, driven by software/tech sector's fast digitization and cost pressures. |
| Augmentation potential | claude-haiku-4-5-20251001 | 4/5 | AI translation and terminology management substantially raise translator productivity by providing drafts, maintaining consistency, and accelerating initial localization, allowing humans to focus on cultural adaptation and quality refinement rather than raw translation. |
| Augmentation potential | claude-sonnet-5 | 5/5 | AI-powered CAT tools, translation memories, and MT substantially speed up translator productivity while humans retain control over cultural nuance and technical accuracy. |
| Task automatability | claude-haiku-4-5-20251001 | 2/5 | While machine translation can produce draft text quickly, adapting software and technical documentation requires cultural context, localization of UI elements, and testing across frameworks—tasks that typically need human oversight and iteration to meet quality standards, falling short of the 50% time-saving threshold for end-to-end automation. |
| Task automatability | claude-sonnet-5 | 3/5 | AI can translate strings and technical docs quickly, but full software localization requires handling context, UI constraints, cultural adaptation, and QA that still need human review to reach equal quality. |
| Adoption barriers | claude-haiku-4-5-20251001 | 3/5 | While no legal requirement mandates human translation, quality standards, liability for incorrect technical documentation, customer expectations for accuracy, and need for human cultural judgment create moderate friction against full automation. |
| Adoption barriers | claude-sonnet-5 | 2/5 | No licensing requirement for translators generally, though some regulated domains (medical/legal software) may require certified human review; otherwise adoption is largely a business/quality decision. |
| Cost vs. human wage | claude-haiku-4-5-20251001 | 2/5 | AI translation plus required human review and cultural adaptation oversight remains labor-intensive and often comparable to hiring skilled translators, particularly for technical and cultural accuracy where errors are costly. |
| Cost vs. human wage | claude-sonnet-5 | 4/5 | MT-assisted localization dramatically cuts per-word costs compared to fully manual translation, though post-editing and cultural QA still add labor costs, keeping it short of the maximum efficiency ratio. |
| Technical feasibility today | claude-haiku-4-5-20251001 | 3/5 | Machine translation APIs and CAT tools with AI assist are deployed in production, but they produce errors in technical terminology, cultural nuance, and contextual accuracy that require human review; no system reliably performs full software/documentation localization without human intervention. |
| Technical feasibility today | claude-sonnet-5 | 3/5 | Machine translation and localization tools (e.g., DeepL, CAT tools with MT integration) are widely deployed in production localization workflows, but typically as part of human-in-the-loop pipelines rather than fully autonomous end-to-end localization. |
Compile information on content and context of information to be translated and on intended audience.
44CI 29–59 · exposure 38 · augmentation 75 · importance 4.2/5 · click for rater detail
Compile information on content and context of information to be translated and on intended audience.
44| Dimension | Model | Rating | Rationale |
|---|---|---|---|
| Sector adoption velocity | claude-haiku-4-5-20251001 | 2/5 | Translation agencies and professional interpreters have adopted AI tools slowly for this preparatory task; most workflows still rely on human research and client interviews. Adoption remains in the pilot phase rather than production-wide displacement seen in higher-digitization sectors. |
| Sector adoption velocity | claude-sonnet-5 | 3/5 | Language services firms are adopting AI-assisted workflows including CAT tools with AI context extraction, but full automation of audience/context analysis is still emerging rather than deeply embedded. |
| Augmentation potential | claude-haiku-4-5-20251001 | 3/5 | AI can assist translators by extracting metadata, suggesting audience profiles, or flagging potential contextual issues, reducing manual research time and improving consistency, though the translator typically refines and validates these suggestions before use. |
| Augmentation potential | claude-sonnet-5 | 5/5 | AI significantly speeds up gathering background information, terminology, and audience context, letting translators focus on higher-order interpretive decisions. |
| Task automatability | claude-haiku-4-5-20251001 | 2/5 | While AI can extract basic textual content and identify metadata, compiling nuanced understanding of context and audience intent requires domain expertise and real-time interaction that current systems handle inconsistently. Full automation would need to reliably capture implicit cultural, technical, and communicative context that often requires human judgment. |
| Task automatability | claude-sonnet-5 | 3/5 | AI can gather and summarize contextual information (domain, register, audience) from provided materials or web research, but synthesizing nuanced client intent and audience-specific requirements still typically needs human judgment and clarification dialogue. |
| Adoption barriers | claude-haiku-4-5-20251001 | 4/5 | Professional translation standards, industry regulations in regulated sectors (legal, medical, financial), and client accountability for accuracy create strong expectations that qualified humans compile context; errors in context assessment directly affect translation quality and liability, limiting automation substitution. |
| Adoption barriers | claude-sonnet-5 | 2/5 | No licensing requirement for this preparatory task, though quality/liability concerns in specialized domains (legal, medical) create some incentive for human review. |
| Cost vs. human wage | claude-haiku-4-5-20251001 | 3/5 | AI tools for content analysis and audience profiling have relatively low inference costs, but they require meaningful human oversight, validation, and integration into translation workflows, offsetting savings and bringing overall cost closer to hiring a human research assistant. |
| Cost vs. human wage | claude-sonnet-5 | 4/5 | AI can quickly scan and summarize source content and audience context at a fraction of the cost of a human analyst spending time on preparatory research. |
| Technical feasibility today | claude-haiku-4-5-20251001 | 2/5 | No deployed product reliably performs comprehensive context and audience compilation for translation workflows. AI tools can tag some metadata or suggest audience segments, but interpreters and translators still manually verify and refine context understanding in production settings, indicating material gaps in current systems. |
| Technical feasibility today | claude-sonnet-5 | 3/5 | Translation platforms and LLM-based tools can extract glossary terms, domain context, and stylistic cues from source documents, but this is usually a supporting feature rather than a standalone reliable product function performed at scale. |
Check original texts or confer with authors to ensure that translations retain the content, meaning, and feeling of the original material.
36CI 30–41 · exposure 25 · augmentation 75 · importance 4.3/5 · click for rater detail
Check original texts or confer with authors to ensure that translations retain the content, meaning, and feeling of the original material.
36| Dimension | Model | Rating | Rationale |
|---|---|---|---|
| Sector adoption velocity | claude-haiku-4-5-20251001 | 2/5 | Translation and localization sectors are adopting AI for draft generation, but verification and QA remain heavily human-driven in professional contexts. Adoption of full automation for quality assurance is slow due to perceived risk of meaning loss in high-stakes content. |
| Sector adoption velocity | claude-sonnet-5 | 3/5 | Translation and localization industries have adopted AI/MT tools broadly, but the specific quality-assurance and author-consultation step lags behind adoption of raw translation automation. |
| Augmentation potential | claude-haiku-4-5-20251001 | 4/5 | AI translation systems and comparison tools can meaningfully assist human translators and editors by generating candidate translations and flagging potential inconsistencies, allowing the human expert to focus verification effort on meaning and tone rather than mechanical accuracy. |
| Augmentation potential | claude-sonnet-5 | 4/5 | AI can efficiently highlight potential meaning shifts, suggest alternative renderings, and speed up the reviewer's comparison process, meaningfully boosting productivity while the human retains final judgment. |
| Task automatability | claude-haiku-4-5-20251001 | 2/5 | While AI translation systems can generate draft translations quickly, verifying that content, meaning, and feeling are retained requires nuanced judgment about cultural idiom, tone, and authorial intent that current systems struggle with consistently. The task involves comparison, judgment calls, and author consultation—elements that still require substantial human oversight. |
| Task automatability | claude-sonnet-5 | 2/5 | AI can compare translations for fidelity and flag discrepancies, but confirming that nuanced meaning and 'feeling' are preserved and conferring with authors requires human judgment and interpersonal interaction that current systems cannot fully replace. |
| Adoption barriers | claude-haiku-4-5-20251001 | 3/5 | There are some organizational and professional norms favoring human translators and editor review, but no hard legal barrier prevents automated verification tools. Clients and publishers may prefer human sign-off, creating friction rather than regulatory prohibition. |
| Adoption barriers | claude-sonnet-5 | 2/5 | No licensing requirement typically applies, but client/publisher expectations for human quality assurance and accountability for meaning-critical documents (literary, legal, medical) create moderate friction. |
| Cost vs. human wage | claude-haiku-4-5-20251001 | 2/5 | AI translation tools are cheap, but verification and refinement still requires skilled human translators or subject-matter experts to review and consult with authors, making the total cost of an automated-then-verified pipeline comparable to or more expensive than direct human translation. |
| Cost vs. human wage | claude-sonnet-5 | 3/5 | AI-assisted comparison is cheap for bulk text, but the conferring-with-authors component still requires human time, keeping overall cost roughly comparable once verification labor is included. |
| Technical feasibility today | claude-haiku-4-5-20251001 | 2/5 | No deployed product reliably performs end-to-end verification of translation fidelity, meaning preservation, and emotional tone at scale. Translation QA tools exist but typically flag mechanical errors rather than assessing whether feeling and deeper meaning are captured, which remains largely a human editorial function. |
| Technical feasibility today | claude-sonnet-5 | 2/5 | Some translation QA tools flag inconsistencies or literal mismatches, but no deployed product reliably assesses stylistic/emotional fidelity or conducts author consultations in production workflows. |
Identify and resolve conflicts related to the meanings of words, concepts, practices, or behaviors.
31CI 25–36 · exposure 25 · augmentation 63 · importance 4.3/5 · click for rater detail
Identify and resolve conflicts related to the meanings of words, concepts, practices, or behaviors.
31| Dimension | Model | Rating | Rationale |
|---|---|---|---|
| Sector adoption velocity | claude-haiku-4-5-20251001 | 2/5 | Adoption of AI for interpretation remains limited to supporting roles (terminology lookup, draft translation, accessibility captioning). Core conflict resolution in professional interpreting is still predominantly human-driven; uptake of autonomous AI systems for this task is slow and concentrated in low-stakes or high-volume domains. |
| Sector adoption velocity | claude-sonnet-5 | 3/5 | Translation is an early and fast-adopting AI use case, but the specific sub-task of resolving deep meaning conflicts remains a mostly manual, expert-driven activity even in fast-adopting firms. |
| Augmentation potential | claude-haiku-4-5-20251001 | 3/5 | AI translation and terminology tools meaningfully assist human interpreters by providing quick reference options, flagging ambiguous terms, and generating draft renderings that interpreters can critique and refine. This augmentation improves speed and consistency without removing human judgment from conflict resolution. |
| Augmentation potential | claude-sonnet-5 | 4/5 | AI tools substantially speed up drafting and flag potential ambiguities or mistranslations, letting human translators focus judgment on resolving genuine conceptual conflicts. |
| Task automatability | claude-haiku-4-5-20251001 | 2/5 | While AI can flag potential semantic ambiguities and suggest alternative interpretations using large language models, resolving cultural and contextual conflicts requires nuanced judgment about speaker intent and cultural practices that current systems handle unreliably. The task demands understanding of pragmatic and cultural context that AI cannot consistently master without human validation. |
| Task automatability | claude-sonnet-5 | 2/5 | This requires nuanced cultural, contextual, and pragmatic judgment to resolve ambiguity or conflicting meanings, which current AI handles inconsistently, especially for idiomatic, culturally-loaded, or high-stakes content.' |
| Adoption barriers | claude-haiku-4-5-20251001 | 4/5 | Professional interpretation and translation are often regulated or credentialed fields; liability for misinterpretation can be substantial (legal, medical, diplomatic contexts), and end-users (courts, hospitals, government) typically require a human translator's accountability and signature. These create strong structural barriers to full automation. |
| Adoption barriers | claude-sonnet-5 | 3/5 | Certified/legal/medical interpretation often requires human accountability and cultural competence, though many everyday translation tasks lack formal licensing barriers. |
| Cost vs. human wage | claude-haiku-4-5-20251001 | 2/5 | Current AI tools (machine translation, terminology databases) reduce some preparation work but do not eliminate the need for expert human interpreters to resolve conflicts. The cost of AI plus required expert oversight approaches or exceeds the cost of direct human interpretation for nuanced conflict resolution. |
| Cost vs. human wage | claude-sonnet-5 | 3/5 | AI translation is cheap per word, but the conflict-resolution portion still requires human expert review, keeping blended cost closer to parity for high-stakes work. |
| Technical feasibility today | claude-haiku-4-5-20251001 | 2/5 | No deployed product reliably resolves semantic and cultural conflicts autonomously in professional interpretation/translation contexts. Tools can assist by highlighting ambiguities, but production systems still require human interpreters to make final judgments on meaning and cultural equivalence, limiting end-to-end automation. |
| Technical feasibility today | claude-sonnet-5 | 2/5 | Machine translation tools exist and flag some ambiguities, but no deployed product reliably identifies and resolves deep semantic/cultural conflicts without human review, especially in professional or legal contexts. |
Discuss translation requirements with clients and determine any fees to be charged for services provided.
30CI 25–35 · exposure 25 · augmentation 63 · importance 3.5/5 · click for rater detail
Discuss translation requirements with clients and determine any fees to be charged for services provided.
30| Dimension | Model | Rating | Rationale |
|---|---|---|---|
| Sector adoption velocity | claude-haiku-4-5-20251001 | 2/5 | Translation and interpretation remain relationship-driven, small-firm-dominated sectors with limited digitization; while agencies use AI for internal triage, live client discussions remain overwhelmingly human-conducted and adoption of autonomous AI for this step is minimal. |
| Sector adoption velocity | claude-sonnet-5 | 2/5 | Translation services sector is adopting AI mainly for the translation itself, with slower uptake of AI for client intake and pricing negotiation processes. |
| Augmentation potential | claude-haiku-4-5-20251001 | 3/5 | AI can assist by drafting fee schedules, summarizing client emails, suggesting pricing models, or preparing project scope templates; a human translator reviewing AI-generated talking points or cost frameworks could work faster, though the core negotiation still requires human presence. |
| Augmentation potential | claude-sonnet-5 | 4/5 | AI tools can help estimate word counts, complexity, and suggest pricing benchmarks, and draft client communications, meaningfully speeding up the human's preparation for these discussions. |
| Task automatability | claude-haiku-4-5-20251001 | 2/5 | While AI can draft translation cost estimates and summarize requirements from text, the negotiation and discussion components—clarifying nuanced client needs, understanding context-specific constraints, and adjusting scope dynamically—require human judgment and relationship management that current systems cannot reliably replicate end-to-end. |
| Task automatability | claude-sonnet-5 | 2/5 | This is a client-facing negotiation and scoping conversation requiring relationship management and business judgment, not just language conversion, which current AI cannot fully replace end-to-end.'},'feasibility' is separate but here it stands alone as low automatability score. Ok I need to output correctly.","rating":2}, |
| Adoption barriers | claude-haiku-4-5-20251001 | 4/5 | Clients typically expect to discuss sensitive translation projects and fee structures directly with a qualified human translator they can trust; regulatory requirements in some jurisdictions (e.g., certified translation) add implicit barriers, and liability for misquoting or misunderstanding scope falls on the translator. |
| Adoption barriers | claude-sonnet-5 | 2/5 | No licensing requirement, but client preference for human rapport and trust in pricing negotiations creates moderate organizational friction. |
| Cost vs. human wage | claude-haiku-4-5-20251001 | 2/5 | The cost of AI-driven client communication tools (setup, integration, error handling, and human oversight of inappropriate responses) is still comparable to or exceeds the cost of a human translator having a brief scoping conversation, especially when relationship continuity matters. |
| Cost vs. human wage | claude-sonnet-5 | 2/5 | Automated quoting tools are cheap but still require human oversight for nuanced negotiations and edge cases, so overall cost savings versus a human handling client relations are modest. |
| Technical feasibility today | claude-haiku-4-5-20251001 | 2/5 | No deployed products reliably perform client-facing discussions and fee negotiation autonomously; AI chatbots can collect basic information but lack the contextual reasoning and interpersonal dexterity needed to handle complex discussions about specialized translation work, rush fees, or custom requirements. |
| Technical feasibility today | claude-sonnet-5 | 2/5 | Some AI chatbot intake tools and quoting software exist for translation agencies, but reliable autonomous negotiation of scope and pricing with clients is not a mature deployed product. |
Adapt translations to students' cognitive and grade levels, collaborating with educational team members as necessary.
29CI 25–32 · exposure 25 · augmentation 75 · importance 4.4/5 · click for rater detail
Adapt translations to students' cognitive and grade levels, collaborating with educational team members as necessary.
29| Dimension | Model | Rating | Rationale |
|---|---|---|---|
| Sector adoption velocity | claude-haiku-4-5-20251001 | 2/5 | K–12 and higher-education institutions adopt automation slowly; educational budgets are constrained, staff are unionized in many jurisdictions, and schools have high organizational inertia. While some use translation tools, systematic AI deployment for student-level adaptation in educational settings remains rare and mostly piloted rather than production-at-scale. |
| Sector adoption velocity | claude-sonnet-5 | 2/5 | Education sector adoption of AI for individualized student services is slow and cautious, especially where student welfare and compliance (e.g., IEPs) are involved. |
| Augmentation potential | claude-haiku-4-5-20251001 | 4/5 | AI translation and readability-level suggestions can significantly assist human interpreters and translators by generating draft adaptations, flagging potential comprehension issues, and proposing grade-level vocabulary alternatives. A translator can then review and refine these suggestions much faster than creating adapted translations from scratch, boosting productivity while maintaining professional control. |
| Augmentation potential | claude-sonnet-5 | 4/5 | AI can meaningfully speed up drafting simplified or adapted translations for interpreters to then review and tailor to a student's needs, serving as a strong assistive tool even if not a full replacement. |
| Task automatability | claude-haiku-4-5-20251001 | 2/5 | While AI can produce initial translations, adapting content to specific cognitive and grade levels requires understanding of individual student profiles, pedagogical context, and nuanced judgment about comprehension barriers. Current systems lack reliable access to student-specific data and cannot consistently match cognitive scaffolding to diverse learner needs without substantial human oversight. |
| Task automatability | claude-sonnet-5 | 2/5 | This requires nuanced judgment about a specific student's developmental level, prior knowledge, and IEP/educational context plus real collaboration with teachers, which current AI cannot reliably assess or execute end-to-end.wa Machine translation can draft simplified text, but adapting it appropriately and coordinating with a team is largely human work. |
| Adoption barriers | claude-haiku-4-5-20251001 | 4/5 | Schools typically require qualified interpreters/translators and employ collaborative decision-making with educational teams; there is often contractual and policy expectation that a certified human professional participates. Liability for educational appropriateness and student outcomes creates organizational friction against full automation, and many districts have credentialing or union agreements protecting interpreter roles. |
| Adoption barriers | claude-sonnet-5 | 4/5 | This task often occurs in special education/ESL contexts requiring qualified, sometimes credentialed interpreters and educator collaboration, with high liability if a student's needs are misjudged, creating meaningful institutional and legal barriers. |
| Cost vs. human wage | claude-haiku-4-5-20251001 | 4/5 | AI inference cost for translation and basic grade-level adjustment is very low compared to the fully-loaded cost of a trained interpreter or translator who customizes for student needs. Even accounting for human review, the AI cost base is at least several times cheaper per output unit. |
| Cost vs. human wage | claude-sonnet-5 | 2/5 | AI translation is cheap, but the human oversight, pedagogical judgment, and collaborative meetings needed to properly adapt content for a specific student keep overall costs comparable to or only modestly below human-only work. |
| Technical feasibility today | claude-haiku-4-5-20251001 | 2/5 | AI translation tools exist and can produce text, but no deployed product reliably adapts translations to specific cognitive or grade levels without human intervention. Educational applications require validation that translation choices actually match target learners' comprehension, which production systems do not consistently demonstrate. |
| Technical feasibility today | claude-sonnet-5 | 2/5 | Products exist for text simplification and translation, but no deployed system reliably tailors translations to individual students' cognitive/grade levels while integrating with educational teams in production settings. |
Travel with or guide tourists who speak another language.
22CI 14–30 · exposure 17 · augmentation 63 · importance 3.5/5 · click for rater detail
Travel with or guide tourists who speak another language.
22| Dimension | Model | Rating | Rationale |
|---|---|---|---|
| Sector adoption velocity | claude-haiku-4-5-20251001 | 1/5 | Tourism and hospitality remain labor-intensive, in-person sectors with slow AI adoption; most reliance is on human guides, and current AI tools (translation apps) augment rather than replace the guide role. |
| Sector adoption velocity | claude-sonnet-5 | 2/5 | Tourism and hospitality sectors are slower AI adopters compared to information/finance industries, with physical guiding services still overwhelmingly human-delivered. |
| Augmentation potential | claude-haiku-4-5-20251001 | 3/5 | Translation tools and AI-powered guidebooks can assist interpreters by providing quick phrase references or real-time translation support, modestly improving delivery speed and accuracy on the interpretation component. |
| Augmentation potential | claude-sonnet-5 | 4/5 | AI translation tools significantly augment interpreters/guides by handling real-time language conversion, allowing the human to focus on navigation, safety, and personalized engagement. |
| Task automatability | claude-haiku-4-5-20251001 | 1/5 | This task requires real-time, context-aware interpretation while physically accompanying tourists, navigating dynamic social interactions, and making judgment calls about cultural appropriateness—capabilities that current AI systems cannot perform autonomously in embodied form. |
| Task automatability | claude-sonnet-5 | 2/5 | While real-time speech translation apps exist, physically traveling with and guiding tourists requires embodied presence, situational judgment, safety management, and cultural navigation that current AI cannot replicate end-to-end.atibility |
| Adoption barriers | claude-haiku-4-5-20251001 | 4/5 | Tourism and hospitality sectors often have implicit or explicit customer preference for human interaction and personal connection; liability concerns for tourist safety and guidance, plus the embodied and social nature of the task, create significant friction against automation. |
| Adoption barriers | claude-sonnet-5 | 3/5 | No licensing barrier universally exists, but liability for tourist safety, need for real-world judgment in unpredictable environments, and customer preference for human warmth create meaningful friction. |
| Cost vs. human wage | claude-haiku-4-5-20251001 | 2/5 | Machine translation is cheap, but the full task of traveling with and guiding tourists requires physical presence and adaptive decision-making; the cost of any autonomous system capable of this would exceed or match the loaded wage of human interpreters/guides. |
| Cost vs. human wage | claude-sonnet-5 | 2/5 | A human guide-interpreter combines multiple functions (safety, logistics, translation, cultural context) that would require assembling several AI/hardware systems plus human oversight, often costing more than a single local guide in many markets. |
| Technical feasibility today | claude-haiku-4-5-20251001 | 2/5 | While machine translation has improved significantly, no deployed AI system can reliably serve as a physical tour guide or travel companion that handles unexpected conversations, group dynamics, and real-world problem-solving at the quality humans provide; products exist for translation only, not the full embodied task. |
| Technical feasibility today | claude-sonnet-5 | 2/5 | Products like AI translation earpieces and apps assist communication but no deployed product autonomously guides tourists through physical spaces, handles logistics, or manages group dynamics. |
Educate students, parents, staff, and teachers about the roles and functions of educational interpreters.
18CI 11–25 · exposure 8 · augmentation 50 · importance 4.0/5 · click for rater detail
Educate students, parents, staff, and teachers about the roles and functions of educational interpreters.
18| Dimension | Model | Rating | Rationale |
|---|---|---|---|
| Sector adoption velocity | claude-haiku-4-5-20251001 | 2/5 | Educational institutions operate in laggard sectors for AI adoption of interpersonal roles. While schools may use AI for content drafting, the core task of educating and building relationships with stakeholders remains heavily human-centered with slow displacement. |
| Sector adoption velocity | claude-sonnet-5 | 2/5 | Education sector, especially specialized interpreter services, shows slow AI adoption for interpersonal advocacy and training functions. |
| Augmentation potential | claude-haiku-4-5-20251001 | 3/5 | AI can assist by generating educational materials, outlines, and talking points that interpreters and educators can customize and deliver, improving preparation efficiency. However, the core interpersonal and trust-building aspects remain fundamentally human-driven. |
| Augmentation potential | claude-sonnet-5 | 3/5 | AI can help draft educational materials, FAQs, or presentations explaining interpreter roles, aiding preparation even though delivery remains human-led. |
| Task automatability | claude-haiku-4-5-20251001 | 1/5 | This task requires nuanced communication, interpersonal skills, and the ability to tailor educational content to diverse audiences with different backgrounds and understanding levels. Current AI systems cannot reliably conduct live educational sessions or handle the dynamic, context-sensitive dialogue needed to address questions and build trust. |
| Task automatability | claude-sonnet-5 | 2/5 | This is an interpersonal educational/advocacy task requiring live explanation, relationship-building, and context-sensitive communication that AI cannot fully replicate end-to-end today.: |
| Adoption barriers | claude-haiku-4-5-20251001 | 4/5 | Schools and educational institutions typically require qualified human educators and interpreters to deliver training and outreach to build trust, authority, and accountability with parents, students, and staff. Liability and regulatory expectations around educational quality create organizational and professional barriers to substitution. |
| Adoption barriers | claude-sonnet-5 | 3/5 | No strict licensing barrier for this specific task, but organizational norms and the need for trusted human liaison with school communities create moderate friction. |
| Cost vs. human wage | claude-haiku-4-5-20251001 | 2/5 | AI-generated educational materials have low per-unit cost, but the task requires live engagement and relationship-building that demand human presence. Organizations would still need staff to deliver the education, limiting cost savings to supplementary content generation rather than full task replacement. |
| Cost vs. human wage | claude-sonnet-5 | 2/5 | While AI could generate informational materials cheaply, the actual delivery and trust-building with parents/staff/students still requires human presence, limiting cost savings. |
| Technical feasibility today | claude-haiku-4-5-20251001 | 1/5 | No deployed product reliably performs this task end-to-end. While AI can generate generic informational content, educational institutions require personalized, contextually appropriate explanations delivered by humans who can respond to immediate feedback and establish credibility with stakeholders. |
| Technical feasibility today | claude-sonnet-5 | 1/5 | No deployed AI product performs stakeholder education and role clarification for interpreters in schools; this remains a human relational function. |
Train and supervise other translators or interpreters.
11CI 5–16 · exposure 0 · augmentation 50 · importance 3.8/5 · click for rater detail
Train and supervise other translators or interpreters.
11| Dimension | Model | Rating | Rationale |
|---|---|---|---|
| Sector adoption velocity | claude-haiku-4-5-20251001 | 1/5 | Translation and interpretation services are still heavily reliant on human expertise hierarchies; adoption of AI for supervisory roles remains minimal and limited to specialized domains, reflecting slow organizational change. |
| Sector adoption velocity | claude-sonnet-5 | 2/5 | Language services firms are adopting AI for translation quality checks and training aids, but supervisory/management roles remain human-led with slow structural change. |
| Augmentation potential | claude-haiku-4-5-20251001 | 2/5 | AI could assist with mundane documentation, scheduling, or preliminary performance metrics, but the core interpersonal and evaluative aspects of training and supervision leave limited room for meaningful augmentation. |
| Augmentation potential | claude-sonnet-5 | 4/5 | AI can meaningfully assist trainers by generating exercises, evaluating translation quality, and providing feedback data, boosting supervisor productivity substantially. |
| Task automatability | claude-haiku-4-5-20251001 | 1/5 | Training and supervising other translators requires high-level pedagogical judgment, performance assessment, personalized feedback, and real-time decision-making about coaching—capabilities far beyond current AI systems today. |
| Task automatability | claude-sonnet-5 | 1/5 | Training and supervising people is a management and mentorship function requiring judgment, feedback, and interpersonal leadership that current AI cannot perform end-to-end.' |
| Adoption barriers | claude-haiku-4-5-20251001 | 4/5 | Organizations assign training and supervision duties to senior human staff for liability, quality assurance, and personnel management reasons; human judgment and accountability are typically required, creating organizational and fiduciary barriers. |
| Adoption barriers | claude-sonnet-5 | 3/5 | No formal licensing barrier for supervision itself, but organizational structures and accountability for staff performance create real friction against full AI substitution. |
| Cost vs. human wage | claude-haiku-4-5-20251001 | 1/5 | An AI system attempting this task would still require human oversight, quality review, and intervention, making the all-in cost equal to or exceeding the cost of direct human supervision. |
| Cost vs. human wage | claude-sonnet-5 | 2/5 | AI can cheaply generate training content, but the actual supervisory function still requires a human manager, so overall cost savings for the full task are limited. |
| Technical feasibility today | claude-haiku-4-5-20251001 | 1/5 | No deployed products reliably perform end-to-end training and supervision of human translators; this requires sustained interpersonal interaction, assessment calibration, and accountability that AI systems do not perform in production. |
| Technical feasibility today | claude-sonnet-5 | 1/5 | No deployed product manages or supervises human translator/interpreter teams; at best AI provides training materials or practice exercises, not supervision itself. |
Follow ethical codes that protect the confidentiality of information.
0CI 0–0 · exposure 0 · augmentation 13 · importance 4.9/5 · click for rater detail
Follow ethical codes that protect the confidentiality of information.
0| Dimension | Model | Rating | Rationale |
|---|---|---|---|
| Sector adoption velocity | claude-haiku-4-5-20251001 | 1/5 | Confidentiality obligations are immutable regulatory and professional requirements that do not see sector-wide AI adoption; they apply uniformly across interpreting and translation work and cannot be automated away. |
| Sector adoption velocity | claude-sonnet-5 | 1/5 | This is not a task sector adoption data tracks as being automated; it remains a human ethical/legal obligation unaffected by AI tool adoption trends. |
| Augmentation potential | claude-haiku-4-5-20251001 | 1/5 | AI cannot meaningfully assist a human in following ethical codes or confidentiality rules; these are binary compliance obligations that the human professional must own fully and cannot offload or be augmented around. |
| Augmentation potential | claude-sonnet-5 | 2/5 | AI tools with confidentiality safeguards (e.g., secure, non-logging translation systems) can support compliant workflows, but the ethical adherence itself remains a human responsibility. |
| Task automatability | claude-haiku-4-5-20251001 | 1/5 | This task is fundamentally about adhering to ethical and legal obligations regarding confidentiality, which requires human judgment, accountability, and legal responsibility. AI systems cannot autonomously follow professional ethics codes or bear liability for breaches. |
| Task automatability | claude-sonnet-5 | 1/5 | Adhering to an ethical code of conduct is a professional/behavioral commitment requiring judgment, accountability, and trust, not a discrete work output that AI can perform or substitute for.' |
| Adoption barriers | claude-haiku-4-5-20251001 | 5/5 | Professional codes of conduct, licensing requirements, and legal liability create hard barriers: a licensed interpreter or translator must personally uphold confidentiality obligations and is legally accountable for breaches. This cannot be delegated to AI. |
| Adoption barriers | claude-sonnet-5 | 5/5 | Confidentiality obligations are often legally or contractually mandated (e.g., certified interpreter codes, HIPAA, legal privilege), requiring a human bound by professional and legal accountability. |
| Cost vs. human wage | claude-haiku-4-5-20251001 | 1/5 | This task has no direct cost-saving potential for AI; it is a prerequisite compliance obligation that falls on the human professional regardless of whether AI assists. The human must bear the legal and ethical responsibility. |
| Cost vs. human wage | claude-sonnet-5 | 1/5 | There is no AI equivalent performing this compliance/ethics task, so no meaningful cost comparison exists; a human professional must bear this responsibility regardless of AI use. |
| Technical feasibility today | claude-haiku-4-5-20251001 | 1/5 | No deployed product can reliably perform the social, legal, and ethical obligations embedded in professional confidentiality codes. This requires human decision-making and professional accountability that AI cannot assume. |
| Technical feasibility today | claude-sonnet-5 | 1/5 | No deployed product 'follows an ethical code'; confidentiality compliance is enforced through human professional standards, contracts, and legal accountability, not AI systems. |
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.