In the field of technical translation, accuracy and consistency are key, and the most important factor for achieving this is the systematic use of appropriate terminology. Companies that handle technical content, such as user manuals, product specifications, or repair and maintenance instructions, rely on rigorous terminology management to ensure their messaging is clear and consistent across markets.

The centralised control of terminology database management is essential for achieving uniform terminology in all published content, regardless of the volume or complexity of the document. For those responsible for marketing or technical offices, this results in consistent communication, conveyed using the terminology that makes the company stand out, thus avoiding costly errors that could negatively affect the company's image.

Terminology is also key when it comes to improving the outcome of neural machine translation (NMT). One of the main issues affecting the quality and reliability of machine translation (MT) is incorrect terminology that appears when MT engines have not been configured with the appropriate glossaries. The growing demand for glossaries for machine translation requires greater efficiency in the management of these terminology repositories.

AI is gaining ground as an essential tool for supporting the process of creating, maintaining and approving terms.

Firstly, AI can help in extracting terminology from an existing corpus. Before being imported into the dictionary, proposed entries will always be validated by human experts.

Figure 1: AI-powered bilingual terminology extraction.

Artificial intelligence can also assist in checking and correcting terminology once the translation is complete, for both machine and human translations. In the case of machine translation, this process may be especially necessary if a translation has not been done using an engine configured with a subject-specific dictionary.

One of the great advantages offered by AI is its ability to learn and adapt. Through machine learning, tools can suggest specific terms based on the customer's previous preferences, keeping dictionaries up to date with new concepts as they emerge. In addition, AI makes it possible to automate the terminology validation process, reducing manual intervention and minimising human errors. Candidate terms in the target language are automatically recognised from content that has already been translated. Terminology experts can select/confirm proposals and add them to the list of terms. Even definitions in the source language can be generated by the large language model (LLM).

For companies with a global presence, proper management of technical terminology is not only essential for ensuring clear communication about their products and services, but also for protecting their brand and reputation. At STAR, we understand the importance of maintaining consistent terminology and how AI can be a very powerful tool to achieve this.

To find out how we can help you, please contact us.