The April 2026 EUATC knowledge-sharing webinar, delivered in collaboration with ISO TC 37/SC 5, provided an update on how the two cornerstone standards for the language services industry, ISO 17100 for translation services and ISO 18587 for machine translation post-editing, are evolving.
For language service companies (LSCs), the message delivered by SC5 International Chair and 17100 Project Co-Leader Maryse Benhoff and 18587 Project Leader and 17100 Project Co-Leader Livia Florensa is clear: both standards are undergoing significant transformation to reflect current operational realities, particularly the integration of AI-driven workflows and hybrid production environments.
At a structural level, the standards sit within a broader ecosystem under ISO TC 37/SC 5 Working Group 1, built around the concept of translation project specifications and providing a framework that links planning, production, and evaluation across standards such as ISO 11669 (project specification) and ISO 5060 (quality evaluation). This shift signals a move away from static process definitions towards a more dynamic, specification-driven approach to service delivery.
ISO 18587 Machine Translation Post-Editing to be Published End of 2026
The most immediate development concerns ISO 18587 for machine translation post-editing (MTPE). Originally published in 2017, the standard entered systematic review in 2023 and is now approaching completion. The revised version of ISO 18587 is expected to be published by the end of 2026. Crucially, its scope has been expanded beyond traditional machine translation to encompass “non-human translation output”, explicitly including artificial intelligence and large language model (LLM) systems.
This change reflects the technological shift since the standard’s original publication. The updated ISO 18587 will no longer be limited to full post-editing workflows; instead, it acknowledges a wider spectrum of post-editing approaches and introduces hybrid scenarios where translators alternate between human translation, translation memory matches, and AI-generated output within the same task.
Another key development is structural alignment with ISO 17100. The revised 18587 adopts a similar framework for production and post-production processes, preparing the ground for a longer-term consolidation of standards. Additional annexes and harmonised terminology further reinforce consistency across the ISO ecosystem.
ISO 17100 Translation Services to be Published in 2029
While ISO 18587 is nearing publication, the revision of ISO 17100 is at an earlier stage. The current version, published in 2015, defines the core processes and resource requirements for translation services, including mandatory revision and competence criteria. The new revision aims to modernise this framework rather than replace it outright.
On-going work on ISO 17100 is expected to take approximately two to three years, with publication anticipated around 2029. The longer timeline reflects both the complexity of the revision and its strategic importance as the industry’s foundational standard.
The central ambition for ISO 17100 is to integrate the content of ISO 18587, effectively creating a unified standard covering all translation workflows: human, machine-assisted, and hybrid. However, this integration will be gradual. As highlighted during the webinar, the consolidation effort is expected to span several years, with ISO 18587 continuing to exist independently in the interim.
In terms of content, several major themes are shaping the revision of ISO 17100. First, there is a strong emphasis on risk management. Future versions are expected to incorporate more explicit risk assessment mechanisms, enabling translation service providers (TSPs) and clients to determine appropriate workflows based on the intended use and criticality of the content.
Second, the standard will remain technology-neutral. Rather than prescribing specific tools or solutions, it will describe processes and competencies that can accommodate evolving technologies, including AI-assisted quality assurance and automated workflows. This is intended to prevent the standard from becoming obsolete in a rapidly changing technological environment.
Third, there is a shift from role-based definitions (translator, reviser, post-editor) to people- and competence-based frameworks. The revised standard will therefore focus on competencies such as revision, post-editing, and evaluation, rather than fixed job titles.
A further area of active discussion focuses on risk assessment, transparency, and disclosure. While assessing and managing risk has been identified as a key element to include, the working groups are considering whether, and to what extent, TSPs should disclose the use of non-human translation in their processes. While no final decision has been reached, the topic is closely linked to broader regulatory developments, including the EU AI Act, and to client expectations around accountability and quality assurance.
Practical Implications for Language Service Companies
For LSCs, the practical implications are relatively measured in the short term. Both standards will be published with transition periods, allowing certified organisations to adapt gradually during their normal recertification cycles. More importantly, the revisions are designed to make the standards easier to apply in real-world workflows, reducing the gap between formal compliance and operational practice.
In strategic terms, however, the direction of travel is significant. The convergence of ISO 17100 and ISO 18587, combined with the introduction of translation project specifications and risk-based approaches, signals a move towards a more flexible, outcome-oriented model of standardisation. Rather than prescribing fixed processes, future standards will define frameworks within which different workflows (human, automated, or hybrid) can be deployed according to client needs.
The timeline is therefore twofold: a near-term update with the revised ISO 18587 expected in 2026, and a longer-term transformation culminating in a unified ISO 17100 around 2029.
Together, these developments mark the most substantial evolution of translation standards in over a decade, aligning them more closely with the realities of an AI-enabled language services industry.
