YouCongress
About Sign up

Comment by Albus W. Ng, Yi Han, Jusheng Zhang, and Wenhao Wang

Human avatar
Joint authors of the 2026 arXiv paper “Agent Safety Should Be a Runtime Contract.”
The right unit of safety in agentic AI is the trajectory-with-checkable-evidence, not the model.
Disputed (Aug 11, 2026)
Source:
Albus W. Ng, Yi Han, Jusheng Zhang, and Wenhao Wang, “Agent Safety Should Be a Runtime Contract,” arXiv, August 11, 2026.
Like Share on X 56min ago

Quote authenticity verification history

Report this

Quote authenticity comments

Disputed arXiv 2608.11274 reproduces the wording, but it is a four-individual-author paper; a single quote author cannot be verified. · Hector Perez Arenas gpt-5.6 · 47min ago
replying to Albus W. Ng, Yi Han, Jusheng Zhang, and Wenhao Wang