📈 AI Efficiency, 😎 Human Quality — Accelerating Translation for Billions of Words 🌍
Track: Advanced Localization Management | AL9 |
Friday, June 7, 2024, 3:45pm – 4:30pm
Held in: Herbert room
Presenters:
Adam Bittlingmayer - ModelFront
Conchita Laguardia - FARFETCH
Moderator: Johan Sporre
Host: Donna Parrish
High-volume buyers are accelerating the translation of high-value content by integrating AI that automatically confirms as many segments as safely possible. Trillions of translations are looked at by human translators but left untouched. Machine translation is now so good that professional human translators often spend most of their time and cognitive effort confirming segments without editing a single character. However, the industry failed to turn this amazing AI progress into amazing efficiency. GenAI started in translation in 2017, but after half a decade, in most localization workflows, humans were still forced to review every single word of machine translation fully manually. Now AI that learns to automatically confirm segments is available and seamlessly integrated into translation management systems. This session will cover the practical aspects of rolling out AI for accelerated post-editing, based on concrete experience inside a translation team processing billions of words.
Takeaways:
- What is the new accelerated workflow?
- How much more efficient is it than fully manual post-editing?
- How is quality compared to fully manual post-editing?
- How do these AI models integrate into existing tools, workflows, and teams?
- What are common mistakes and keys to success?
