Four years into the GenAI era, every language services company has had to answer the…
Welcome to the ATC’s first-ever quarterly newsletter: Spotlight on AI. Each quarter, we’ll be bringing you the latest AI news, opinion and best practice from across the UK and global language services industry, curated by the ATC team, member companies, and AI innovators and experts.
We’ll be covering everything from new tools and technology to governance, ethics, quality and compliance.
For this first edition, we’ve collated a selection of articles looking at the current state of AI legislation and share an update on emerging ISO standards for AI. Alongside this, in each edition you’ll hear from Mark Jones, Head of Technology at Comtec Translations, who’ll share with us his thoughts on what’s hot in the realm of AI right now.
Enjoy!
EU Commission’s Free AI Language Tools
The European Commission’s free and secure AI language tools allow you to translate, generate, and improve content, building on the professional translation expertise and databases and leveraging the power of AI.
EU AI Act: Compliance Checker
The 2024 EU AI Act harmonised rules on Artificial Intelligence, and set out a risk-based approach to the development and deployment of AI technologies. The European Commission’s Single Information Platform’s Compliance Checker is a great place to start understanding how the regulations may impact your business.
What About UK Legislation?
The UK does not have any AI-specific regulation or legislation covering AI as a technology. Instead, AI is regulated in the context in which it is used, through existing legal frameworks, such as financial services legislation. Some regulators have oversight of the development, implementation and use of AI more broadly. For example, the Information Commissioner’s Office has guidance on its website covering AI and data protection.
Emerging ISO Standards for AI
Increasingly, AI conversations with clients focus on understanding, managing, and mitigating the risks associated with the safe and secure use of AI for sensitive business content. A whole host of ISO standards for AI is emerging to respond to a changing legislative and regulatory landscape.
AI Localization Think Tank: Hear it from the Experts
The AI Localization Think Tank is a collective of AI and localization experts, innovators, and problem-solvers from across the globe whose mission it is to make sense of the AI evolution through informed, diverse, and inclusive perspectives. Dive in, read the analyses and thought pieces, and consume the curated resources.
Musings By Mark: Customisation is Key
Customisation, orchestration, and governance are likely to be three of the most recurring themes we hear this year when it comes to the application of AI in translation workflows. While all are equally important, customisation is enabling small-to-medium LSCs to stay ahead of the curve.
For a long time, training and building custom Neural Machine Translation, or NMT, models seemed out of reach for many small and medium-sized LSCs due to the volume of data, resources, time, and expertise required. The arrival of LLMs has changed this paradigm and levelled the playing field, enabling more granular, lightweight customisation layers through prompt engineering and terminology.
These two elements have become a focus for many teams, but they require expertise and guidance from language professionals to make a real impact. Knowing how to handle terminology for an LLM may differ from the requirements of a detailed terminology database used in legacy workflows. This, alongside prompting techniques that require rigorous testing and, in some cases, failure before success, allows vendors of all sizes to offer brand- and client-specific MT at a much lower entry point.
This not only improves the client experience but also enables our linguist teams to make a more significant impact in these hybrid workflows. It removes the need to correct repetitive, frustrating issues and allows linguists to focus instead on cultural and stylistic refinement.
This shift across much of our industry is reflected in the launch of DeepL’s Customization Hub in November last year, which gives even less technical or non-localisation users the ability to influence how their MT performs.
Other advanced techniques, from translation RAG with vector stores to knowledge graphs and agentic workflows, are becoming more prevalent and offer huge value for more advanced teams. However, the value of lightweight customisation remains a key entry point for anyone looking to stay ahead of the curve.
Mark Jones, Head of Technology at Comtec Translations
Related Posts
