To promote exchanges and collaboration between Taiwan and France in the field of AI evaluation, the AI Testing Laboratory of the Center for Measurement Standards (CMS), Measurement Technology Development Center, Industrial Technology Research Institute (ITRI), has maintained ongoing engagement with France's National Laboratory of Metrology and Testing (LNE). In December 2025, the LNE team visited Taiwan for engineer-level technical exchanges. Both sides shared views on trends and practical experience in language model evaluation and discussed next steps for collaboration, including advancing a Memorandum of Understanding (MOU), to further deepen the Taiwan–France partnership.
Engineer-Level Exchanges Focused on Practical Experience in Language Model Evaluation
This exchange aimed to build a shared understanding and strengthen mutual trust for future collaboration. Discussions focused on core concepts and methodologies of language model evaluation, covering areas such as test item bank development, metric design thinking, quality principles, and data governance. Through face-to-face dialogue, both parties compared their respective practices and experience, identified differences in implementation, and further accumulated a foundation for scalable collaboration. Key topics included:
- Exchanges on evaluation dimensions and approaches to metric design for language model assessment
- Sharing concepts for test bank development and quality control (including requirements such as consistency and reproducibility)
- Discussions on principles for test data governance and usage policies
- Sharing experience on presenting evaluation results and principles for report writing
Advancing the MOU and Strengthening a Long-Term Collaboration Mechanism
To ensure continuity and sustained progress, both sides also discussed collaboration approaches under the MOU framework, including future exchange formats, coordination cadence, and potential themes for joint work. The AI Testing Laboratory will continue regular engagement with LNE, gradually expanding the depth of Taiwan–France collaboration and international linkages in language model evaluation. Through accumulated experience in cross-border collaboration, Taiwan’s professional capacity and international visibility in AI evaluation will be further strengthened.
Strengthening International Alignment to Support Trustworthy AI Development
Amid global momentum toward trustworthy AI and growing attention to international norms and compliance, cross-border practical exchanges help enhance the transparency and comparability of evaluation methodologies, while aligning perspectives on testing and quality principles. Building on the existing collaboration foundation, the AI Testing Laboratory will continue promoting international cooperation and professional exchanges, supporting Taiwan's AI industry and application stakeholders with more comprehensive evaluation references and technical support as they engage with international markets and compliance requirements.
About LNE
The National Laboratory of Metrology and Testing (Laboratoire national de métrologie et d'essais, LNE) is France's official institution for metrology and standards. LNE has long been dedicated to testing technologies and quality assurance and continues to participate in AI evaluation and standardization-related work, collaborating with international partners to advance trustworthy AI evaluation methods and cooperation mechanisms.
