On Safety Risks in Experience-Driven Self-Evolving Agents
Fuente:
arXiv
Saved in:
| Main Authors: | , , , , , , , , , , |
|---|---|
| Format: | Preprint |
| Published: |
2026
|
| Subjects: | |
| Online Access: | |
| Tags: |
Add Tag
No Tags, Be the first to tag this record!
|
| _version_ | 1866911604192313344 |
|---|---|
| author | Zhao, Weixiang Zhang, Yichen Wang, Yingshuo Deng, Yang Zhao, Yanyan Zhi, Xuda Huang, Yongbo HaoHe Che, Wanxiang Qin, Bing Liu, Ting |
| author_facet | Zhao, Weixiang Zhang, Yichen Wang, Yingshuo Deng, Yang Zhao, Yanyan Zhi, Xuda Huang, Yongbo HaoHe Che, Wanxiang Qin, Bing Liu, Ting |
| contents | Experience-driven self-evolution has emerged as a promising paradigm for improving the autonomy of large language model agents, yet its reliance on self-curated experience introduces underexplored safety risks. In this study, we investigate how experience accumulation and utilization in self-evolving agents affect safety performance across web-based and embodied environments. Notably, experience gathered solely from benign tasks can still compromise safety in high-risk scenarios. Further analysis attributes this degradation to the execution-oriented nature of accumulated experience, which reinforces agents' tendency to act rather than refuse. In more realistic settings where agents encounter both benign and harmful tasks, refusal-related experience mitigates safety decline but induces over-refusal, revealing a fundamental safety-utility trade-off. Overall, our findings expose inherent limitations of current self-evolving agents and call for more principled strategies to ensure safe and reliable adaptation. |
| format | Preprint |
| id |
arxiv_https___arxiv_org_abs_2604_16968 |
| institution | arXiv |
| publishDate | 2026 |
| record_format | arxiv |
| spellingShingle | On Safety Risks in Experience-Driven Self-Evolving Agents Zhao, Weixiang Zhang, Yichen Wang, Yingshuo Deng, Yang Zhao, Yanyan Zhi, Xuda Huang, Yongbo HaoHe Che, Wanxiang Qin, Bing Liu, Ting Computation and Language Experience-driven self-evolution has emerged as a promising paradigm for improving the autonomy of large language model agents, yet its reliance on self-curated experience introduces underexplored safety risks. In this study, we investigate how experience accumulation and utilization in self-evolving agents affect safety performance across web-based and embodied environments. Notably, experience gathered solely from benign tasks can still compromise safety in high-risk scenarios. Further analysis attributes this degradation to the execution-oriented nature of accumulated experience, which reinforces agents' tendency to act rather than refuse. In more realistic settings where agents encounter both benign and harmful tasks, refusal-related experience mitigates safety decline but induces over-refusal, revealing a fundamental safety-utility trade-off. Overall, our findings expose inherent limitations of current self-evolving agents and call for more principled strategies to ensure safe and reliable adaptation. |
| title | On Safety Risks in Experience-Driven Self-Evolving Agents |
| topic | Computation and Language |
| url | https://arxiv.org/abs/2604.16968 |