senckađ
Group745
Group745
Group745
Group745
Group745
Group745
Thought Leaders in association withPartners in Crime
Group745

Dr. Strange Localisation: Or, How I Learned to Stop Worrying and Love AI

05/06/2023
Production Company
London, UK
306
Share
Tanya Bogin, MD of Craft London on AI’s impact on the localisation industry

The process of adapting marketing campaigns for a global audience is evolving rapidly. Workflows are being turned on their heads, automation is a basic reality rather than a new feature, and generative content platforms are quickly saturating our consciousness and our laptops. In all this change, I’ve been considering the wider impact of the phenomenon of one type of AI, the large language model, on the creative localisation industry. 

As part of an army of people in global production whose job is to spearhead innovation and build teams that can translate new technologies into client solutions, I’m not unaccustomed to major paradigm shifts in global marketing. That said, very occasionally an industry disruptor hits the scene that sends even me for a loop. 


Should we be worried about AI?

When the new world of generative AI and large language models (LLMs) hit the scene in a big way, the debate moved beyond: ‘how can we adapt marketing content to other languages with AI and is it good?’ into something more philosophical. With the proliferation of easy-to-engage-with LLMs like Chat GPT and Bard, has the time come to take the “trans” out of transcreation? Is localisation dead? Are we all just creators now? 

AI language models are a new dimension of global content creation and just like machine translation, they can also adapt content to other languages. However, don’t be fooled. For now, Chat GPT and Bard are even less appropriate for transcreation than neural machine translation. Even if data protection, privacy and secure deployment wasn’t a concern (and they are), LLMs don’t produce high degrees of quality, accuracy, and integrity with the original source (especially considering brand guardianship, global consistency, regulatory requirements). In certain languages the results produce fluent “semi-translations”, but it depends on the language.  

The true paradigm-shifting potential of LLMs for language adaptation are more interesting. Our R&D around its cultural applications is quickly changing how we support our writers in achieving better transcreation. Using it as a cultural research tool can speed up the process of concept validation in multiple markets. It can breathe life into source content and explain wordplay to otherwise non-native writers (in their local language). It has use cases in preparing content for audiovisual localisation and integration with AI-synthetic voice (and face and body for that matter) workflows. It can transform how quickly we architect client toolkits like brand glossaries and style guides. 


AI’s impact on the localisation industry 

Large language models will have a deep impact on the industry, and roles will inherently change. 

An AI-specialised copyeditor (formerly transcreator or copywriter) may soon be editing AI generated local content derived from a simple prompt given by a “linguistic prompt engineer” (formerly localisation engineer). 

In 2016, my team set up a global newsroom of copywriters and designers from 14 countries in an “agile content factory” to cover the Rio Olympics for a client. Today a skilled writer/prompt expert could generate 14 markets worth of original content in an instant. But there’s a catch. You would still require 14 (now remote) markets worth of creative language specialists, to decode, validate, and ensure brand integrity before going live. 

The lesson: Roles will evolve, not disappear. 


We need not fear new, only adapt. 

When all is said and done, is Chat GPT the end of human-based transcreation? Is the language industry done for? No, I really don’t think so.  

I feel a responsibility to prove this out. Not only to my clients, whose brand voice must be creatively sound and globally consistent yet culturally appropriate, but also to the thousands of native speaking writers, transcreation specialists, and translators in my extended team. 

If for example, we can simulate local cultural environments to help our writers and operators produce work better, faster, and more creatively in 200+ countries, then we can produce exponentially more native content more efficiently. More content means more work for language specialists – at first translators, then transcreators - now learning to add global content creation and AI-editing to their repertoires. Harnessing that power for a purpose-driven brand that wants to make the world a better, more sustainable, more diversity-nurturing place? 

Maybe AI isn’t really that scary after all. 



Tanya Bogin is MD of Craft London 

Credits
Work from Craft London
Liberty
Nespresso
13/02/2024
5
0
Path to Pro - Keep Inspiring
Sky Broadband
19/12/2023
8
0
Path to Pro - Fortnite Winner
Sky Broadband
19/12/2023
11
0
ALL THEIR WORK