Responsible Agentic AI Requires Explicit Provenance
The paper discusses the need for explicit provenance in agentic AI to enhance public trust and accountability. It argues that current frameworks lack the necessary mechanisms to assign responsibility when harm occurs. The authors propose a structured approach to establish and formalize provenance throughout the AI lifecycle.
- ▪Agentic AI is increasingly used in various domains, yet public trust is lagging.
- ▪The authors emphasize that explicit provenance is essential for making responsibility computable and actionable.
- ▪They outline a framework that includes identifying responsibility gaps and formalizing what provenance must encode.
Opening excerpt (first ~120 words) tap to expand
Computer Science > Artificial Intelligence arXiv:2605.17169 (cs) [Submitted on 16 May 2026] Title:Responsible Agentic AI Requires Explicit Provenance Authors:Jinwei Hu, Xinmiao Huang, Qisong He, Youcheng Sun, Yi Dong, Xiaowei Huang View a PDF of the paper titled Responsible Agentic AI Requires Explicit Provenance, by Jinwei Hu and 5 other authors View PDF HTML (experimental) Abstract:Agentic AI is rapidly proliferating across diverse real-world domains such as software engineering, yet public trust has not kept pace.
…
Excerpt limited to ~120 words for fair-use compliance. The full article is at arXiv cs.AI.
Discussion
0 commentsMore from arXiv cs.AI
