ISO 18587 and Machine Translation Post-Editing (MTPE): Balancing Speed and Human Quality
As neural machine translation becomes a standard component of enterprise localization, ISO 18587 provides the international benchmark for full and light post-editing workflows, defining translator competencies and quality verification standards.
Distinguishing Full Post-Editing from Light Post-Editing under ISO 18587
With neural machine translation (NMT) reaching high baseline fluency across major language pairs, enterprise localization buyers increasingly request Machine Translation Post-Editing (MTPE) to accelerate turnaround times and manage high-volume content. Published by the International Organization for Standardization, ISO 18587 (‘Translation services — Post-editing of machine translation output — Requirements’) establishes the formal governance framework for human post-editing. The standard clearly distinguishes between ‘Light Post-Editing’—aimed solely at making internal or perishable content comprehensible and accurate—and ‘Full Post-Editing’, which requires output to be stylistically and terminologically indistinguishable from human translation from scratch.
Post-Editor Qualifications and Pre-Translation Risk Assessment
Crucially, ISO 18587 stipulates that post-editors must hold the exact same professional qualifications as human translators under ISO 17100: a recognized degree in translation or linguistics, or equivalent documented domain experience. Because NMT engines produce fluent-sounding sentences that can mask subtle factual omissions, mistranslated negations, or terminology inconsistencies, certified LSPs combine pre-translation termbase injection with rigorous human line-by-line verification against the source text.
Executive Takeaways for Global Enterprises
- Key Insight: ISO 18587 defines strict process requirements for Full Post-Editing (human-equivalent publishable quality) vs. Light Post-Editing (gist comprehension).
- Key Insight: Under ISO 18587, post-editors must meet the same linguistic and subject-matter qualification criteria required by ISO 17100.
- Key Insight: Neural MT requires heightened vigilance against ‘fluent mistranslations’ where plausible target syntax conceals source-text inaccuracies.
¿Necesita traducciones multilingües certificadas, traducciones juradas o localización conforme a las normas ISO para sus proyectos internacionales? Solicite un presupuesto gratuito personalizado a Alphatrad y reciba una respuesta detallada en 24 horas.
Editorial & Copyright Attribution Note: This industry analysis has been independently synthesized and adapted by Alphatrad’s editorial team based on verified regulatory documentation and industry research originally published by ISO – International Organization for Standardization. For the original publication and primary documentation, please consult:
Source: https://www.iso.org/standard/62970.html
