To promote international exchange and collaboration in AI evaluation, the AI Testing Laboratory of the Center for Measurement Standards at ITRI welcomed a delegation from UL Solutions on March 12, 2026. The visiting team included key contributors involved in the development of UL 3115, along with technical representatives. During the visit, both sides exchanged views on AI evaluation mechanisms, large language model testing procedures, alignment with international standards, and potential directions for future collaboration. They also held preliminary discussions on possible cooperation topics and next-step approaches, laying a foundation for continued partnership.
Visit Focused on AI Evaluation Mechanisms and Testing Practices
The visit aimed to build mutual understanding and strengthen trust for future collaboration. Discussions focused on the current operational framework of the AI Testing Laboratory, evaluation procedures for large language models, and the generation of evaluation reports, while also introducing Taiwan's progress in advancing its AI evaluation system. Through presentations, a laboratory tour, and in-depth discussions, both sides gained a better understanding of each other's plans and approaches in AI evaluation, testing methodologies, and system development. Key topics of discussion included:
- An introduction to the AI Testing Laboratory, the AIEC evaluation framework, testing procedures, and service offerings
- Exchanges on the five key evaluation dimensions for large language models and the related reporting mechanism
- Discussions on UL 3115, the laboratory's AI evaluation methods, and directions for alignment with international standards
- Preliminary discussions on future collaboration topics, demonstration cases, and exchange formats
Exchange on Future Collaboration Directions and Continued Follow-up
During the meeting, UL Solutions expressed strong interest in the laboratory's AI evaluation mechanisms, operational model, and approaches to future application development. Both parties also exchanged views on possible areas of collaboration, including alignment of evaluation methodologies, technical exchange, promotion of industry collaboration, and potential cooperation frameworks. Relevant collaboration details are still under discussion, and both sides will continue to advance concrete cooperation plans based on future progress in these discussions, further strengthening the connection between Taiwan's AI evaluation efforts and international institutions.
Strengthening International Alignment to Support the Development of Trustworthy AI
As global attention to trustworthy AI, AI safety, and international compliance continues to grow, engagement with international testing, inspection, verification, and standards-related organizations helps Taiwan stay informed of global trends and strengthen the alignment of its AI evaluation mechanisms with international frameworks. The AI Testing Laboratory will continue to promote international cooperation and professional exchange, enhance its evaluation methods, testing procedures, and technical service capabilities, and support domestic industries and application sectors in addressing international market and compliance requirements.
About UL Solutions
UL Solutions is an international organization engaged in testing, inspection, and verification. It has long been committed to product safety, compliance verification, and technical services, while also maintaining close attention to AI safety, risk management, and related testing and evaluation issues. The organization actively engages in exchange and collaboration with industry and technical partners.
