ChatGPT in Localization

Watch the keynote from the conference that broke industry records for live attendance.

Rome – March 9, 2023

We're pleased to share the video of our CEO Marco Trombetti's opening keynote presentation at the ChatGPT in Localization event organized by More than 2,500 individuals from the industry were drawn to the event, and it set a new record, with 1,300 people watching simultaneously. The talk addresses the impact of ChatGPT – an "unreasonably good next-word predictor" created by our friends at OpenAI – and other large language models on the translation and localization industry.

The keynote compares ChatGPT with machine translation, starting with their shared origins in a Transformer model.

The presentation illustrates the 3 possible scenarios we are about to face and their impact on the localization industry: content generation in the target language, content restructuring, and large-scale content creation.

Overall, the keynote provides valuable insights into the future of language translation, offering a glimpse into cutting-edge technologies and techniques that will drive our industry forward.

What is ChatGPT?

ChatGPT is a chatbot based on a large language model developed by OpenAI. An LLM is a deep learning algorithm that can recognize, summarize, translate, predict, and generate text based on knowledge gained from massive datasets. ChatGPT’s primary function is to engage in conversation with users and provide helpful responses to their questions. It went from 0 to 100 million users in a record time of 2 months.

ChatGPT Translation Quality: How Good Is It at Translating?

While not designed to process translations, ChatGPT appears to be pretty good at it. Its output is fluent but not as accurate as required. The keynote highlights the many similarities and key differences between ChatGPT and machine translation, from Google Translate to adaptive neural MT like ModernMT. The presentation also explores ChatGPT alternatives – including GPT-3, BLOOM, Megatron Turing, Chinchilla, LLaMA – and discusses why LLMs could soon become outdated as we move towards brain-to-computer communication interfaces.

Impact of Large Language Models on Translators, Copywriters, and Language Service Providers

During his talk, our CEO also provided insights into how large language models such as ChatGPT will affect the roles of translators, copywriters, and language service providers. For every one of the 3 possible scenarios described above, an analysis is provided of how translation will evolve and who will play a crucial role. Finally, the keynote addresses the implications of a future scenario in which we consume content exclusively within a chatbot-like app, marking the end of the traditional "Internet deal", i.e. publishing content to drive traffic to a website.

All the people working on machine translation are in the right place at the right time to have an impact on the future of humanity. We believe in the capacity the language community has to invent new solutions to allow everyone to understand and be understood in their own language.
Marco Trombetti – Translated CEO