Generative AI loves humans in the loop

(Written by AI & edited by Frederik R. Pedersen, Founder and CEO, generative AI company EasyTranslate)


Generative AI is going to change the world, but humans in the loop are going to be vital if it is going to live up to its potential. It will transform translation and localisation.

When businesses expand into new markets and target audiences who speak different languages, the need for effective communication is paramount. Traditionally, translation has been the way to bridge linguistic gaps and to do it in the most ‘local’ way possible. 

As US writer Marshall McLulan once said the medium is indeed the message, but that message has to deliver the same meaning, whatever language it comes in. That’s not always an easy thing to do. Great translation and localisation depends on a number of factors.  

For instance, a Danish Christmas celebration may differ significantly from an American one and a direct translation may not adequately convey those differences. As a result, content may come across as inauthentic or culturally insensitive, potentially alienating customers and hindering market penetration.

Translation is a complex discipline that often fails to capture the nuances of local culture and context, but that is changing rapidly for the better. Generative AI technology is changing translation to the extent that translation-as-we-know-it is becoming redundant in today’s global market.

Translation is dead, long live New Translation, but what does this mean? 

With the advent of generative AI, the potential for fine-tuning models on a company level is stratospheric. The opportunity to seamlessly communicate with any community in any language and to know that all parties will thoroughly understand that message is Utopian, but it is in our grasp.

Contrary to media speculation, this will not mean the end for humans in the process. While generative AI is going to disrupt translation forever, it will be humans that will take the discipline to another level.

Recently, the Times in London reported that Sue Brooks, Managing Director of the Reuters News Agency would be ‘astonished’ if generative AI did not report on news stories, especially financial results, by the end of 2023.

She went on to say that as the company introduced a raft of AI features to its services, there was ‘always a human in the loop’ to ensure an optimal service.

At EasyTranslate, we have been talking about humans in the loop for a long time. We believe that it is the perfect combination for customers who want to utilise Large Language Models (LLMs) to their optimal advantage.

By bringing in humans to finesse the early work of generative AI, there is likely to be a technological evolution towards proprietary Small Language Models (SLMs), meaning that not only will translation will be perfect from language-to-language, it will also accurately convey the particular tone and language of each customer.

This approach saves clients money because AI, which is cheaper than human capital, does the heavy lifting with the bulk of copy and translations first, before a translator and/or copy editor is brought in for specific parts of the process to improve it. In this process, copy editors also assist to refine machine learning. 

This is a game-changer.

Beyond the cases where people are trying to understand a text merely written in an unknown language, the goal for all translation tasks is communicating with a local audience in their own language with the aim to establish that process as locally relevant.

By bringing together technology and humans in this way, that Utopia can be reached and then taken to an even higher level. This is not pie-in-the-sky thinking about a perfect future, it is happening now and it is happening very quickly.

Translation, while essential for conveying basic information, does not fully resonate with a target audience. By bringing together generative AI and those humans in the loop, it does.

Original content created with the help of AI and combined with human copywriters, offers a more scalable and adaptive solution for businesses looking to expand into new markets. 

Instead of simply translating text, generative AI creates content tailored to the specific cultural context and preferences of the target audience. By fine-tuning the AI models/the system on a company level, businesses  ensure that their messaging is consistent and effective across different regions and languages, without losing the essence of its brand identity.

The rise of generative AI and fine-tuning models may initially seem threatening to the translation industry. However, it’s important to recognise that these technologies will also complement and enhance the work of translators. 

Rather than replacing human expertise, AI-GC can assist translators in generating more accurate and culturally appropriate translations, while also increasing efficiency and reducing turnaround times.

Additionally, the translation industry can adapt by focusing on higher-level tasks that involve a deep understanding of cultural nuances, such as transcreation or localisation consulting. By embracing these new technologies, translators will continue to play a crucial role in facilitating global communication.

While AI has made significant advancements, it still struggles with accurately capturing the intricacies of language, tone, and context. Human translators, with their cultural knowledge and intuition, will continue to be essential in ensuring that content is not only linguistically accurate but also culturally relevant and appealing to the target audience.

We also believe that we need to offer humans the direct chance to work with customers. Our free freelance marketplace connects verified translators and copy editors to localise content and ensure it’s consistent with a brand’s identity and voice.

The future of global communication lies in a collaborative approach that combines AI-generated content and fine-tuning models with the irreplaceable skills and expertise of human translators. 

By working together, businesses can create more effective and culturally sensitive content, ultimately fostering better connections with their target audiences around the world.