#DutyOfCare

The words #DutyOfCare, introduced in the opening remarks and subsequent keynote, became the conference takeaway.

I hope our duty of care towards the viewers, but also towards all our co-workers in our market whose jobs are affected as our sector changes, becomes a focal theme for everyone in 2025!

In the past couple of years, media localization has faced some of its hardest times. For the first time that I can remember, revenues declined. We witnessed scriptwriter, actor and voice talent strikes, so production volumes were affected for a while, as was dubbing in some countries. We’ve seen layoffs across all stakeholders – from content owners, to LSPs, to individual translators complaining about lack of work. Production volumes still haven’t recovered and the uncertainty around AI adoption doesn’t help either, even if an increased number of non-English local productions have taken the world by storm and are paving the way for an “entertainment without borders” that is produced in a plurality of languages and consumed by a plurality of audiences, anywhere at any time. Live shows and events have entered the streaming arena dynamically, turnaround times continue to shrink and cost pressures abound. The demand for access services for people with disabilities, through sign language interpreting, subtitling and captioning, and audio description, is increasing – and it is expected to grow further once the European Accessibility Act is rolled out next year. 

ChatGPT was launched a couple of years ago and suddenly AI became the buzzword on everyone’s lips. Machine learning technologies were here before, but they had not become a household name until LLMs and Gen AI came into the picture. There are all sorts of scenarios about what might happen now. The AI trend has taken the world by storm in all industries and verticals. In the media market, video localization automation solutions are springing up one after the other, with AI dubbing being one of the hottest trends. Many (if not most) content owners are looking for such tech-enabled solutions to allow them to make their expanded catalogues accessible and available into more languages at significantly reduced costs and faster than ever before. 

Technology and MT is of course something that we’ve been exploring in previous editions of our conference, but in the past two years this discussion has come to a new level and brings many ethical questions with it. New automation workflows are being trialled and introduced everywhere, new tech-enabled LSPs are making their appearance, human-in-the-loop workflows are becoming commonplace – or machine-in-the-loop, or expert-in-the-lead if you prefer these terms instead. At the same time, media localization professionals are striking to have their voices heard, issue position statements about the dangers of GenAI in translation, and are concerned about the sustainability of our profession.

While it’s true that audiences now have unprecedented access to localized and accessible content, it can also be argued that quantity has been trumping quality for some time, and it is time we ask ourselves whether ‘good enough’ quality is really good enough for our society. News outlets increasingly rely on AI-generated content, as well as automated captions and subtitles for the deaf and hard-of-hearing. Translation is becoming part of every application that uses an LLM. Slator reports on this trend and calls it Translation as a Feature (TaaF). 

Despite the ongoing improvement of AI in translation, it still presents a number of pitfalls, such as lack of contextual and cultural understanding, difficulty capturing linguistic nuance and the potential for bias. All these issues, unless properly addressed by experts, can result in mistranslation, misinformation, miscommunication… and texts that rather than igniting your imagination, leave you uninspired by their mediocrity.

These are the very texts that our children are growing up with. The language we provide in news items, audio descriptions, subtitled and dubbed films and series, which constitute the “mass literature” of our century, has a proven impact in the learning and literacy of the viewers. Language is the cornerstone of culture, it is what makes us human. Through language we have built civilisations, developed science and medicine, literature and philosophy. If AI language has the potential to impoverish our world cultures, what are the ethical implications we have as an industry? When considering ethics beyond AI, there are growing movements focused on social inclusion and diversity, for which  language has a very important role to play. Which begs the question:

What is our duty of care as an industry?