---
id: 20261017-T0-01
title: "美国90%工作时间花在AI基准测试覆盖不到的任务上"
title_en: "90% of US Work Hours Spent on Tasks AI Benchmarks Don't Cover"
url: https://ai.daily.yangsir.net/daily/20261017-T0-01
issue_date: 2026-10-17
publish_date: 2026-10-16T17:08:38.000Z
category: research
source_name: "HN AI 精选"
source_url: https://www.mercor.com/research/labor-index/benchmark-map/
---

# 美国90%工作时间花在AI基准测试覆盖不到的任务上

Mercor研究发布劳动指数基准地图，数据显示美国90%的工作时间用于执行AI基准测试未能有效覆盖的任务。这意味着当前AI评测体系与实际经济活动之间存在巨大缺口：基准测试高分并不代表AI能胜任大多数真实工作。该研究为AI能力评估和劳动力市场影响分析提供了新的数据视角。

## English Version

**90% of US Work Hours Spent on Tasks AI Benchmarks Don't Cover**

Mercor's Labor Index benchmark map reveals that 90% of US working hours are spent on tasks not well covered by existing AI benchmarks. This exposes a major gap between current AI evaluation frameworks and real economic activity: high benchmark scores do not mean AI can perform most real-world jobs. The research offers a new data perspective for assessing AI capabilities and analyzing labor market impact.

---

**来源**：[HN AI 精选](https://www.mercor.com/research/labor-index/benchmark-map/)

**详情页**：https://ai.daily.yangsir.net/daily/20261017-T0-01

---

*智语观潮 · Daily — https://ai.daily.yangsir.net/llms.txt*