AsianFin -- The world’s first operational safety testing standard for AI agents has been officially released by the World Digital Technology Academy (WDTA) at the United Nations headquarters in Geneva.

Titled the AI STR Series: Operational Safety Testing Standard for AI Agents, the new global benchmark was developed under the leadership of Ant Group, Tsinghua University, and China Telecom. The standard was co-authored by more than 20 institutions and companies, including PwC, Nanyang Technological University in Singapore, and Washington University in St. Louis.

The standard marks a significant milestone as the first of its kind dedicated to testing the safety of individual AI agents. It aims to fill a critical gap in the industry’s technical framework for ensuring secure and reliable AI agent deployment.