# 新研究发现，人工智能审核员可以发现大多数代理不当行为，但会导致审核人员误报大量错误

> 全文为英语

**Published:** 2026年11月7日
**Author:** Manuel Uth
**Source:** The Decoder
**Language:** en-US
**HTML:** https://aiqazaqstan.com/zh/articles/ai-auditors-catch-most-agent-misbehavior-but-flood-reviewers-with-false

根据人类研究员安迪·霍尔的一项研究，人工智能模型通常可以在日志中发现其他人工智能代理的问题行为。 GPT-5 发现了大约 98% 的奖励黑客案例，但搜索调整得越彻底，错误警报就越多。那还没有……

AI models can often spot problematic behavior by other AI agents in logs, according to a study by Anthropic researcher Andy Hall. GPT-5 found about 98 percent of reward-hacking cases, but the more exhaustively the search is tuned, the more false alarms pile up. That's not yet enough for reliable safety checks. The article AI auditors catch most agent misbehavior but flood reviewers with false alarms, new study finds appeared first on The Decoder.

---

*内容来自原始来源： [The Decoder](https://the-decoder.com/ai-auditors-catch-most-agent-misbehavior-but-flood-reviewers-with-false-alarms-new-study-finds/). 查看原文.*

**标签：** #OpenAI #Anthropic

---

*AI Qazaqstan — non-commercial informational portal about AI in Kazakhstan. Machine translations and summaries are AI-generated.*