Interpreters and translators

Use computer-assisted translation tools and machine translation software for efficiency

This task is already heavily automated—CAT tools, translation memories, and MT engines are the standard workflow. AI systems now integrate these tools seamlessly, auto-suggesting translations, managing terminology, and applying previous translations, reducing human effort to acceptance/rejection decisions.

BLS evidence: Nearly all translators use software in their work, including computer-assisted translation (CAT) tools and machine translation software that translators then review in a process called post-editing.

Translate written materials from source language into target language

Neural machine translation (DeepL, GPT-4, etc.) now produces high-quality translations of routine written materials that require only light human review for accuracy. For standard business documents, articles, and correspondence between major languages, AI performs at or above median translator quality with minimal post-editing.

BLS evidence: Translators convert written materials from one language into another language, maintaining or duplicating the written structure and style of the source text while keeping ideas and facts accurate.

Compile information and technical terms into glossaries and terminology databases

AI can extract technical terms from documents, suggest definitions, identify inconsistencies, and populate terminology databases with minimal human input. LLMs excel at pattern recognition across large corpora to build glossaries, though domain experts still validate specialized or ambiguous terms.

BLS evidence: Interpreters and translators compile information and technical terms into glossaries and terminology databases for use in their oral renditions and translations.

Review and revise translations through multiple iterations before finalization

AI can perform multiple revision passes, checking consistency, grammar, terminology adherence, and flagging potential errors across iterations. Modern LLMs can self-critique and refine translations with each pass, though final human approval remains standard practice for quality assurance.

BLS evidence: Translations often go through several revisions before becoming final.

Adapt and localize content including cultural references and expressions for target audiences

AI can now identify cultural references, suggest localized alternatives, and adapt idioms for target audiences with strong accuracy for common language pairs and mainstream content. Humans review for cultural sensitivity and edge cases, but AI handles the bulk of routine localization work autonomously.

BLS evidence: Translators must properly transmit cultural references, including slang, and other expressions that do not translate literally, and localizers engage in adapting text and graphics to make products appear to have originated in the target country.

Manage business operations including pricing, billing, and client relations

AI tools can automate invoicing, track billable hours, manage client databases, and handle routine client communications. Pricing calculations and basic CRM functions are highly automatable, though relationship management and negotiating complex contracts still benefit from human involvement.

BLS evidence: Self-employed interpreters and translators must be able to manage their finances, set prices for their work, bill customers, keep records, and market their services to build their client base.

Convert concepts, style, and tone from source language to equivalent target language

Advanced LLMs can now capture style, tone, and conceptual equivalence across languages with significant success, though complex literary devices, cultural wordplay, and subtle tonal shifts still benefit from human judgment. AI handles most routine style conversion autonomously but humans review for high-stakes or creative content.

BLS evidence: Interpreters and translators typically convert concepts, style, and tone in the source language to equivalent concepts, style, and tone of the target language.

Convert spoken messages from one language into another in real-time or consecutively

Real-time speech interpretation requires low-latency processing, handling of accents/dialects, and managing conversational dynamics that current AI struggles with in live settings. AI can assist with vocabulary but humans must manage the real-time flow and contextual nuance, especially for consecutive interpretation where speaker intent and emphasis matter.

BLS evidence: Interpreters convert information from one spoken language into another—or, in the case of sign language interpreters, between spoken language and sign language.

Perform sight translation of written documents into spoken language

Sight translation requires simultaneous reading comprehension and oral production under time pressure, often in legal/medical settings where accuracy is critical. While AI can translate text-to-speech, the real-time cognitive load of reading unfamiliar documents and speaking fluently remains a distinctly human skill that AI cannot yet replicate reliably.

BLS evidence: Sight translation interpreters provide translation of a written document directly into a spoken language for immediate understanding.

Facilitate communication in specialized settings such as courts, hospitals, or conferences

Specialized settings require physical presence, real-time human judgment in high-stakes scenarios (legal testimony, medical emergencies), and managing interpersonal dynamics that AI cannot replicate. Courts and hospitals will not trust AI for liability reasons, and the human mediator role is essential for clarifying misunderstandings.

BLS evidence: The BLS describes multiple specialized roles including healthcare interpreters who help patients communicate with medical staff, legal interpreters who work in courts, and conference interpreters who work at events with non-English-speaking attendees.