AI News AI资讯 8h ago Updated 2h ago 更新于 2小时前 46

Sprains, pain, and whiplash: Waymo and Zoox test drivers are getting hurt as robotaxis scale 扭伤、疼痛和挥鞭伤:Waymo和Zoox的测试司机在机器人出租车扩张中受伤

Waymo and Zoox test drivers sustained over two dozen injuries in 2024–2025 from hard braking and sudden autonomous vehicle movements, per OSHA data reviewed by TechCrunch Waymo reported 16 injuries across three cities, with incidents escalating from 5 in 2024 to 11 in 2025, primarily caused by hard braking with no apparent obstruction Zoox reported up to 8 injuries linked to hard braking, with contractors confirming "brake jabs" and system "nogo" shutdowns were still occurring as recently as Jul Waymo和Zoox的自动驾驶测试司机在2024-2025年间因车辆急刹车或突然移动导致数十人受伤,包括扭伤、拉伤、挥鞭伤等 急刹车多由自动驾驶系统误判路面障碍物或系统故障(内部代码"nogo")触发,部分案例中车辆在无遮挡情况下突然制动 OSHA数据显示Waymo报告16起相关工伤,Zoox报告至少8起,但数据仅覆盖员工超100人的高危险行业企业 随着Waymo车队从2025年5月的1500辆扩张至3500辆,测试里程增加,工伤事故数量同步上升 Zoox测试人员透露截至2026年7月急刹车问题仍未解决,但公司称此类事件占数百万英里测试里程的比例极小

68
Hot 热度
62
Quality 质量
65
Impact 影响力

Analysis 深度分析

TL;DR

  • Waymo and Zoox test drivers sustained over two dozen injuries in 2024–2025 from hard braking and sudden autonomous vehicle movements, per OSHA data reviewed by TechCrunch
  • Waymo reported 16 injuries across three cities, with incidents escalating from 5 in 2024 to 11 in 2025, primarily caused by hard braking with no apparent obstruction
  • Zoox reported up to 8 injuries linked to hard braking, with contractors confirming "brake jabs" and system "nogo" shutdowns were still occurring as recently as July 2026
  • The injuries—ranging from sprains and whiplash to fractures requiring months away from work—highlight a safety gap in autonomous vehicle testing as fleets rapidly scale
  • OSHA disclosure requirements only capture companies classified as "higher-hazard" industries with 100+ employees, meaning injuries at other AV companies or in smaller markets may be unreported

Why It Matters

This reveals a critical occupational safety concern in the autonomous vehicle industry: as robotaxi fleets scale aggressively, the human test drivers absorbing the risks of unrefined autonomous systems are suffering real, documented injuries. For AI practitioners and robotics engineers, it underscores that validation and testing phases cannot treat human operators as expendable and that autonomous driving software still has significant reliability gaps in real-world edge cases.

Technical Details

  • Waymo's test fleet grew from over 1,500 vehicles in May 2025 to more than 3,500 today, with testing intensity rising alongside expansion into new cities like Phoenix and Los Angeles
  • Zoox test vehicles are modified Toyota Highlander SUVs equipped with its self-driving system; hard-braking incidents ("brake jabs") can occur multiple times per drive, often triggered by the system detecting or mistakenly detecting road debris
  • A "nogo" event—internal code for a full system shutdown—causes abrupt emergency braking at speeds up to 45 mph, throwing unrestrained or loosely restrained occupants forward
  • Injury reports document specific failure modes: exaggerated braking without obstruction, sudden reverse while self-parking with steering wheel whipping, and harsh stops in response to false positives
  • OSHA data covers only establishments with 100+ employees in classified higher-hazard industries, creating a significant reporting blind spot for smaller operations and non-covered AV companies

Industry Insight

  • Rapid fleet scaling without proportional investment in test driver safety protocols and software maturity is creating a preventable occupational health crisis; companies must treat human test operators as a critical safety asset, not a cost to minimize
  • The prevalence of false-positive braking events suggests autonomous perception systems still struggle with debris classification and edge-case decision-making—prioritizing robust simulation and corner-case training could reduce both injuries and public road risks
  • Regulatory bodies should consider expanding OSHA-style injury reporting requirements to all AV testing operations regardless of company size or industry classification, ensuring transparent accountability as the sector scales

TL;DR

  • Waymo和Zoox的自动驾驶测试司机在2024-2025年间因车辆急刹车或突然移动导致数十人受伤,包括扭伤、拉伤、挥鞭伤等
  • 急刹车多由自动驾驶系统误判路面障碍物或系统故障(内部代码"nogo")触发,部分案例中车辆在无遮挡情况下突然制动
  • OSHA数据显示Waymo报告16起相关工伤,Zoox报告至少8起,但数据仅覆盖员工超100人的高危险行业企业
  • 随着Waymo车队从2025年5月的1500辆扩张至3500辆,测试里程增加,工伤事故数量同步上升
  • Zoox测试人员透露截至2026年7月急刹车问题仍未解决,但公司称此类事件占数百万英里测试里程的比例极小

为什么值得看

这篇文章揭示了自动驾驶商业化快速扩张背后被忽视的测试人员安全风险,为行业提供了关于人机协作安全的重要警示。对于AI从业者和自动驾驶企业而言,这提醒了在追求技术验证和规模化部署时,必须将测试人员安全纳入核心考量,而非仅关注里程数据。

技术解析

  • Waymo车队规模从2025年5月的1500辆增长至目前的3500辆,测试覆盖旧金山、洛杉矶、凤凰城等城市,但测试专用车辆数量未单独披露
  • Zoox测试车队使用改装丰田汉兰达SUV,配备自研自动驾驶系统,内部将急刹车称为"brake jab"
  • "nogo"是Zoox内部代码,指整个自动驾驶系统shutdown,常导致车辆在高速行驶中突然急刹,测试人员需在车流中快速接管
  • 急刹车触发原因包括系统误判路面碎片等障碍物,或无明确原因的异常制动,部分案例中单次测试可发生多次急刹
  • OSHA数据仅要求员工超100人的高危险行业企业公开, newer cities的测试站点数据可能存在遗漏

行业启示

  • 自动驾驶企业在快速扩张测试规模时,需建立更完善的人员安全保护机制,包括改进急刹预警系统、优化人机接管流程
  • 当前监管披露机制存在盲区,行业应推动更全面的工伤数据透明化,确保中小规模测试站点的安全问题不被忽视
  • 企业在追求技术验证速度的同时,应将测试人员安全置于与乘客安全同等重要的地位,避免将"数百万英里"的统计数字掩盖个体风险

Disclaimer: The above content is generated by AI and is for reference only. 免责声明:以上内容由 AI 生成,仅供参考。

Autonomous Driving 自动驾驶 Robotics 机器人 Security 安全 Ethics 伦理 Policy 政策