How does the language translation function work in CMS Online?

CMS Online now includes a translation function to help make content more accessible to a global audience. Here are some key points to understand about this feature:

1. Machine Translation: The translations provided are generated by an advanced machine learning language service. While highly accurate, please note that this content has not been audited by human subject matter experts in the target language. This means that while the overall context and meaning of CMS content should be preserved, some nuances or specific terminology might not be perfect.

2. Use of Glossaries: To enhance the accuracy of translated content, we use specialized glossaries that add proper context to industry-specific terms and phrases. Glossaries are essential because they ensure that the most accurate and relevant translations are used, particularly for technical or specialized content. This helps maintain the integrity and clarity of the material, making it easier for readers to understand complex concepts across languages.

3. Quality: We leverage top-tier translation services known for their quality and reliability. These services employ sophisticated algorithms and large datasets to provide language conversions that are both precise and contextually appropriate. These reasons all help ensure that users can confidently engage with CMS Online content in their preferred language.

4. Feedback on Terminology: We recognize that industry terminology can vary, and native speakers might have preferred terms for certain concepts. If you have suggestions or specific terminology that you believe would enhance the translation quality of CMS Online content, please share them with us using the IHRDC Customer Support form. Your feedback is invaluable in helping us improve and tailor the experience to better meet your needs.

5. Privacy:

We use enterprise-grade translation services, including DeepL and Google Translate, under commercial agreements that explicitly prohibit the use of our data for model training.

All content processed through these services is handled in accordance with strict data protection and confidentiality standards. In enterprise configurations, neither DeepL nor Google retains or uses submitted text to improve their models. Your data remains your data and is not incorporated into any shared or public training datasets. Additionally, we do not send any personally identifiable information for translation as it stays within IHRDC's secure enclave.

This approach ensures that sensitive information, proprietary content, and student or customer data are protected at all times while still benefiting from high-quality machine translation.