Reinventing Translation: Ethics, AI, and the Future of Meaning
- Eva Premk Bogataj
- Oct 10
- 9 min read
The Future of Translation Studies, Agencies, and Meaning
“To translate is not to repeat — it is to dream in another rhythm.” — From my studies on translation of Lermontov
From Reflection to Reinvention
Two decades ago, I stood in a quiet classroom, comparing translations of Lermontov’s “I Walk Alone on the Road.”
My aim was purely poetic: to trace how rhythm, silence, and emotion survive migration from one language to another.
What began as a study of verse has become a study of systems.
Today, that same question — how meaning travels — has expanded beyond literature into technology.
Translation, once a bridge between cultures, has become a crossroads between humans and machines.
As artificial intelligence learns to mirror not only language but tone, nuance, and emotion, the translator’s task is itself being translated: from craft to curation, from words to architectures of meaning.
Where We Stand: The 2025 Language Industry
The global language-services market reached USD 71.7 billion in 2024 and is projected to grow to USD 75.7 billion in 2025 — a modest but steady +5.6 % rise. Growth now comes mainly from machine-translation post-editing (MTPE) and multilingual content management.
Yet according to the ELIS 2025 report (1,322 respondents in 50 countries), most freelancers and agencies report negative growth for 2024 and increasingly cautious expectations for 2025. The profession is not shrinking — it is shifting.
AI has already changed the workflow. The share of projects including MTPE rose from 26 % (2022) to 46 % (2024). Quality-evaluation metrics such as COMET / MetricX / XCOMET now rival human raters — but only at the level of form, not intention or style. Post-editing has become the new baseline skill.
Yet the human role moves up the value chain: from typing words to managing meaning, tone, ethics, and consistency.
Switzerland as a Case Study: Multilingualism as Infrastructure
If you want to understand where the future of translation is being quietly shaped, look to Switzerland — a country where language is not merely a medium of expression, but a framework of coexistence.
Here, translation is not a profession on the margins; it is a pillar of governance and social trust.
86 % of Swiss residents consider multilingualism essential for national cohesion, according to the Federal Statistical Office (2024).
Nearly two thirds of the population use at least two languages daily — not only in education and administration, but in families, workplaces, and digital spaces.
Federal institutions and cantons are legally required to communicate in German, French, and Italian, while English is increasingly used for scientific, financial, and international communication.
Since 2023, public authorities have also adopted Plain Language and Accessibility directives, ensuring that official information can be understood by citizens across linguistic and cognitive borders.
In the public procurement framework (rev. 2023), translation and linguistic services are now evaluated by the “most advantageous offer” principle — weighting quality, consistency, and ISO compliance over lowest cost.
Agencies certified under ISO 17100 (translation processes) and ISO 18587 (MT post-editing) receive clear preference, as transparency, security, and terminological consistency become mandatory procurement criteria.
The result is a translation ecosystem that blends ethics, policy, and technology:
Universities and language service providers cooperate in training experts for terminology management, accessibility, and MTPE.
Federal offices are implementing AI-assisted terminology databases (TERMCOORD-CH, eCH-Standards).
Even in the private sector, companies like Swiss Post, UBS, and Roche maintain in-house language teams that ensure multilingual coherence across AI-augmented workflows.
Switzerland demonstrates that in a multilingual democracy, translation is far more than ornament or support.
It is the infrastructure of inclusion, the architecture of comprehension, and — increasingly — the interface between human precision and machine efficiency.
In this sense, the Swiss model points toward the global future of translation: not the disappearance of translators, but their transformation into curators of clarity in an age where language is both data and destiny.
4. The Standards that Define Value
ISO 17100 — defines translation processes, qualifications, and mandatory two-eye review.
ISO 18587 — regulates full post-editing of machine translation, aiming for output “indistinguishable from human.”
Agencies and universities that implement these frameworks signal not only professionalism but accountability.
In a world flooded with instant text, traceability becomes prestige.
5. Technology in 2025: The New Toolbox
Google Translate + Gemini: conversational mode, live translate, and AI-guided pronunciation practice.
DeepL Language AI / Voice: enterprise-grade privacy, real-time speech translation, and API integration.
Large-context LLMs (Gemini 1.5, Claude, GPT-5): handle full-document consistency, tone matching, and contextual nuance.
CAT tools remain essential: Trados, memoQ, Phrase, Wordfast, Smartcat — now embedded with MT and quality-estimation dashboards.
Technology no longer replaces translators — it redefines them.The future translator is a meaning engineer: combining linguistic intuition with metric literacy.
How Translation Schools Can Reinvent Themselves
From Translator to Language Operations Specialist
When I translated my first book from Russian at the age of nineteen, I thought of translation as pure art — a dialogue between two souls across languages. Since then, I have translated across disciplines and worlds: from literature to law, from cultural essays to corporate reports.
By twenty-five, I had the privilege of leading translator training at Slovenia’s second-largest translation company — a role that opened the door to global industry standards, emerging CAT tools, and the first serious debates about machine translation.
Furtheron, working with leading translation agenciy, I deepened this experience — moving from word-level precision to process design, from linguistic craft to communication strategy. These decades have shown me that translation, at its best, is not an isolated act of language transfer but a system of meaning management.
And yet, many translation programs still train as if the world stopped in 2005. Some universities have embraced technology and interdisciplinary learning; others still treat translation as a static craft rather than a dynamic ecosystem.
As AI systems, localization workflows, and multilingual data pipelines transform communication, the question for academia is no longer how to preserve translation — but how to evolve it.
Tomorrow’s translators will not compete with machines; they will design, audit, and curate them. They will bridge technology and ethics, ensuring that automation remains transparent, humane, and meaningful.
To achieve this, translation studies must expand into Language Operations (LangOps) — uniting linguistic precision with data literacy, intercultural insight, and process leadership.
Below is a suggested 1–2-year curriculum for translation schools that want to prepare students for the next decade of global communication:
MTPE & Quality Evaluation: ISO 18587 post-editing processes, automated Quality Evaluation (QE), COMET-based scoring, and hybrid human–machine workflows.
Terminology & Ontology Design: Creation of cross-domain glossaries, ontology mapping for AI systems, and integration into CAT and Translation Management Systems.
Regulated Contexts: Plain-language communication and compliance writing for government, finance, and healthcare — focusing on clarity, inclusion, and accessibility.
Transcreation & Brand Voice: Cultural adaptation, UX microcopy, and multilingual brand storytelling for the digital economy.
AI for Language Professionals: Prompt engineering, ethical model use, bias control, and collaboration with generative AI tools.
Professional Standards & Ethics: Application of ISO 17100 / 18587, GDPR, intellectual property, confidentiality, and sustainability in translation practice.
Practicum — Language Ops Lab (12 weeks): Hands-on collaboration with real clients (NGOs, city councils, or financial institutions) to manage multilingual workflows and quality metrics.
7. How Agencies Can Evolve
From Language Service Providers to Meaning Architects
The next decade will not belong to agencies that translate — but to those that transform.
In a landscape where AI can already render fluent text in seconds, the competitive edge lies not in speed, but in stewardship of meaning — the ability to merge technology, empathy, and ethics into coherent, measurable value.
The translation agency of the future will not sell “words per page,” but clarity per context. Its task will be to make complex information understandable, inclusive, and trustworthy — across languages, media, and cultures.This requires a shift from linear production to circular collaboration: translation as part of a living communication ecosystem.
New Service Lines — From Text to Trust
MTPE as a Service (ISO 18587): Offer structured human-in-the-loop post-editing with SLA-based quality metrics, benchmarking models like COMET, BLEURT, and human preference tests.
Transcreation & Tone Curation: M ove beyond literal adaptation: help brands build multilingual tone-of-voice systems, UX microcopy libraries, and emotional resonance guidelines.
Terminology & Knowledge Design: Develop terminology programs as living ontologies — with version control, client dashboards, and integration into CMS and AI training data.
Multilingual Accessibility & Plain Language: Provide inclusive communication audits for governments, banks, and healthcare institutions, aligning with EU accessibility and ISO 24495-1 (Plain Language).
AI Audit & Quality Evaluation (QE): Introduce ethical and technical QA for machine translation and LLM output — including bias detection, factuality checks, and explainability reports.
New Revenue Models — From Projects to Partnerships
Strategic Retainers — Offer continuous maintenance of terminology, QE, and LLM fine-tuning — positioning your team as long-term language partners, not ad-hoc vendors.
Feature-Based Pricing — Replace per-word models with value metrics: charge for adaptation, optimization, and creative decision-making (microcopy, SEO, UX testing).
Compliance & Governance Packages — Bundle translation with ISO, ESG, and data-integrity audits — turning linguistic expertise into organizational risk management.
New Roles — From Translator to Orchestrator
Editor-Curator — designs the human signature of machine output.
Terminologist & Ontologist — builds structured multilingual knowledge.
MTPE Lead — manages human-AI interaction loops for quality and ethics.
LQA Manager — quantifies tone, coherence, and inclusion across languages.
AI Ops Specialist — ensures transparency, auditability, and alignment.
Agencies that integrate AI + ethics + evaluation will claim the middle ground between automation and artistry —the space where precision meets imagination.
They will become not suppliers of text, but custodians of coherence: ensuring that, even in an age of algorithms, every word still carries the weight of intention.
8. Measuring Meaning — Not Just Output
In the next era of translation, success will not be measured by how much we produce, but by how deeply we connect. Metrics remain essential — but they must evolve from productivity indicators to trust indicators.The following benchmarks combine linguistic precision with human resonance:
Metric | What It Reveals |
COMET / QE Score | Baseline quality benchmark — target within agreed threshold. |
Terminology Consistency (%) | Coherence of brand voice and technical accuracy across touchpoints. |
LQA Error Rate / 1,000 Words | Review discipline and traceability of quality assurance. |
Turnaround Time Reduction (%) | Efficiency gains from MTPE and process optimization. |
Engagement Metrics | Human response — clarity, empathy, and trust in communication. |
Ultimately, translation quality must now include ethical fidelity — alignment with values as much as with vocabulary.
9. Ethics and Responsibility
As translation merges with AI, three duties remain distinctly human.
Protect privacy. Sensitive data must never be used to train public models without consent. Confidentiality is not optional; it is the foundation of trust.
Credit authorship. Transparency about human versus machine contribution is essential. Clients — and readers — have the right to know how meaning was made.
Preserve cultural integrity. Language is never neutral. Bias, erasure, or cultural flattening must be consciously named, measured, and corrected.
Ethics in translation is not about avoiding error — it is about choosing awareness. When translators curate meaning, they also safeguard humanity’s collective voice.
10. The Future of Meaning
Translation will always be more than transfer — it is transformation.
In an age of automation, its purpose expands: to ensure that what is understood remains human.The agencies, educators, and creators who embrace this calling will not compete with machines —they will teach machines what meaning means.
11. What You Can Take Away
If You’re a Translator
Master the new literacy. Learn MTPE (ISO 18587), quality-evaluation metrics (COMET, MetricX, XCOMET), and human-AI collaboration workflows.
Go where machines cannot. Specialize in terminology design, transcreation, or plain-language communication — the domains that require intuition, empathy, and ethical nuance.
Measure your impact. Build a portfolio that shows transformation: before-and-after examples, stylistic refinements, audience reach, and clarity gains.
Stay curious. Engage with AI not as a rival, but as an instrument — one that sharpens your judgment and deepens your authorship.
If You’re an Agency
Evolve from production to curation. Integrate MTPE-driven workflows with human-led tone design and meaning management.
Turn compliance into strategy. Use ISO 17100 and 18587 not just as procedural frameworks, but as arguments for trust, traceability, and premium positioning.
Diversify your offer. Create retainers for terminology, accessibility, and quality-evaluation maintenance — and partner with public, cultural, and financial institutions on inclusive communication.
Lead with responsibility. The future of translation is not repetition — it is stewardship: guiding technology toward ethical, human-centered expression.
Closing Reflection
Translation will always be an act of listening — across languages, systems, and generations.
Its purpose is not to copy the world, but to keep it coherent.
In an age of automation, those who translate with integrity will remain its custodians — guardians of the dialogue that keeps meaning alive.
References & Further Reading
Industry Reports
Nimdzi 100 (2025) — Global Language Services Market Overview.
ELIS 2025 Report (European Language Industry Survey) — Professional outlook, demand shifts, and technology adoption.
CSA Research 2024–2025 — “The Augmented Translator: AI Integration in Practice.”
Standards & Frameworks
ISO 17100 — Translation Services: Requirements for Core Processes and Competence.
ISO 18587 — Post-Editing of Machine Translation Output: Requirements.
ISO 30042:2024 — TermBase eXchange (TBX) for terminology management.
Evaluation & Technology
WMT 2024 / ACL 2025 — Research on COMET, XCOMET and MetricX-24 evaluation models.
DeepL (2025) — Enterprise, Voice, and Contextual AI features.
Google Translate + Gemini (2025) — Conversational and multimodal translation developments.
Policy & Research Context
Swiss Federal Statistical Office (2025) — “Language Use and National Cohesion” (86 % perceive multilingualism as vital).
Swiss Public Procurement Rules (2025) — “Most advantageous offer” and ISO-based quality criteria.
European Commission (2024) — “Language Equality in the Digital Age.”
European Parliament Research Service (2025) — “AI, Translation, and the Future of Multilingualism in the EU.”
Academic & Conceptual Reading
Cronin, Michael (2022). Eco-Translation: Translation and Ecology in the Age of the Anthropocene.
Pym, Anthony (2023). Translation Technology and Ethics.
Venuti, Lawrence (2019). The Translator’s Invisibility: A History of Translation.
Bogataj, Eva P. (2025). When Translation Becomes Interpretation.
