News Center

location: Home > News Center > News > Content

WAIC 2026 | First L1–L5 Tiered Safety Framework for General-Purpose AI Agents

Source:       Time:2026-07-06
Released at the World Artificial Intelligence Conference 2026 by Concordia AI with Shanghai AI Laboratory, AIR and Huawei, the white paper classifies general-purpose agents into five levels of autonomy, judged on two dimensions at once — decision-making autonomy and risk-reduction autonomy — and sets tiered safeguards for each level plus specific constraints for healthcare, finance and industrial control. AIR contributors were Yuanchun Li and Yizhen Yuan.

http://mp.weixin.qq.com/s?__biz=Mzg2OTEzMzExOQ==&mid=2247486799&idx=1&sn=545a669176e3964599531218f8e3a525


上一条:Nature News | Agent Hospital: An All-AI Doctor-Patient Simulation System Empowering Collaborative Care in Real Hospitals 下一条:The Tam Shiu Charitable Trust Supports the Agent Hospital Project

【关闭】

Relevant news

Email:Airoffice@air.tsinghua.edu.cn
Tel:(010)82151160

Address:12 / F, block C, Qidi science and technology building, Tsinghua Science and Technology Park, Haidian District, Beijing

wechat

Jing ICP Bei No. 15006448 | all rights reserved@ Institute of intelligent industry, Tsinghua University