AIST · AI SAFETY / QUALITY MANAGEMENT
Trustworthy AI can fully realize its potential in society when people can use it with confidence. Through AI safety and AI quality management, AIST develops criteria and evaluation technologies, advances international standardization, and promotes social implementation to realize the value of AI in society.
We systematize AI safety and quality management methods and develop guidelines that address the distinctive characteristics of rapidly evolving AI while remaining consistent with knowledge from software engineering and safety engineering.
We develop test datasets, evaluation environments, and evaluation methods as a foundation for assessing diverse aspects of AI, and make them widely available.
We propose the criteria we develop to international standards organizations such as ISO/IEC and advance knowledge originating in Japan into internationally applicable standards, contributing to the adoption of trustworthy AI and stronger industrial competitiveness.
Through guideline explanations, case studies, consortia, and networking events, we encourage the formation of communities working toward trustworthy AI.
Explore by type of result
