Translation, Automation and the STC

Earlier this year, CETRA had the opportunity to contribute a case study to a presentation on the integration of automation in the language industry to the Society for Technical Communication (STC). The presentation was part of the STC’s monthly roundtable, which provides professional development opportunities for technical writers. The topic that month was localization.

The overview of the STC roundtable from the host stated:

The localization industry grew out of the need to make technology products globally accessible. This month will be all about the implications of working with global audiences. My world, the professional translator’s world, is all about facilitating communication between people. Translation is not a commodity. In a global economy, the world becomes more integrated, but technology cannot successfully produce all of the communication humans need and want. Translators are a junction point. This month we will discuss from the basic – defining globalization vs. internationalization vs. localization – and the small – how to write for translation and localization – to the large – what machines are already doing for us and where we see the industry in 20 years.

Key highlights of CETRA’s presentation included:

Machine Translation Is Growing, but Not Disruptive

Language Service Providers (LSP) are increasingly utilizing machine translation (MT) and offering post-edited MT as a separate service. Similarly, freelance translators are also increasingly utilizing MT to support their work.

In fact, according to the Common Sense Advisory’s State of the Market research:

  • MT post-editing was about 8% of the services purchased by clients from LSPs in 2019
  • This compares with 60% of human translation purchased as a service
  • About 36% of LSPs offer post-edited MT as a service in 2019

While neural machine translation has potential for greater optimization, claims of equal quality do not reflect real-world performance. In short, neural MT is not going to replace human translators in the foreseeable future.

Automation has Its Place

Use of MT and automation becomes a function of risk management. Automation has greater use in less complex, more general content that presents lower risk if inaccurate. The consumer and manufacturing industries fit this model.

Conversely, automation has less use in more complex, less predictable content that has greater risk if inaccurate. Examples of this type of content include the legal and medical device industries.

For more, please see our previous blog with the translation automation map.

As noted in the STC roundtable, translators sit at the junction of the use of technology and human creativity. CETRA’s Dominika Weston and Jiri Stejskal attended this year’s annual American Translators Association conference in October (read highlights). At the conference, Dominika noted in one of the sessions she attended that MT should not be viewed as a quality concern. Instead, if used in the correct setting, it can be a benefit to the translation process.

For more information about this topic, please contact info@cetra.com.