← Back to incidents

Amazon Warehouse AI Productivity Tracking Led to Unsafe Working Conditions and Automated Terminations

High

Amazon's AI-powered warehouse productivity tracking system automatically terminated workers who couldn't meet algorithm-set quotas, leading to unsafe working speeds and injury rates significantly higher than industry averages.

Category
Safety Failure
Industry
Technology
Status
Ongoing
Date Occurred
Jan 1, 2019
Date Reported
Apr 29, 2019
Jurisdiction
US
AI Provider
Other/Unknown
Application Type
embedded
Harm Type
physical
Estimated Cost
$50,000,000
People Affected
300,000
Human Review in Place
No
Litigation Filed
Yes
Litigation Status
pending
Regulatory Body
OSHA
Fine Amount
$60,269
workplace_safetyautomated_terminationproductivity_trackingwarehouse_operationslabor_rightsOSHA_violations

Full Description

Amazon deployed AI-powered productivity tracking systems across its warehouse network starting around 2019, using algorithms to monitor worker performance and automatically generate termination notices for employees who failed to meet productivity targets. The system tracked metrics like items picked per hour, with quotas often set based on the performance of top workers without consideration for safety implications or individual worker capabilities. Investigative reporting by Reveal News in April 2019 exposed how this automated system created a culture of speed over safety. Workers reported feeling pressured to work at dangerous speeds to avoid automatic termination, leading to shortcuts that increased injury risk. The Strategic Organizing Center's analysis found that Amazon warehouse workers experienced serious injuries at rates nearly double the industry average, with facilities using AI tracking showing particularly elevated rates. OSHA investigations resulted in multiple citations against Amazon facilities, including a $60,269 fine for exposing workers to ergonomic hazards at a Connecticut facility. The agency found that Amazon's productivity expectations contributed to musculoskeletal disorders and other workplace injuries. Internal Amazon documents revealed the company was aware of the correlation between productivity pressure and injury rates but continued the automated termination practices. The incident affected hundreds of thousands of Amazon warehouse workers across multiple states. Workers reported being terminated after failing to meet quotas due to factors beyond their control, including equipment malfunctions or safety-related slowdowns. The automated system provided little opportunity for human review or appeals, with termination decisions often made without manager oversight. Legislative responses included proposed bills in multiple states requiring disclosure of AI-driven workplace monitoring and quotas. Several class action lawsuits were filed alleging unsafe working conditions and wrongful termination related to the AI tracking systems. The incident drew scrutiny from labor rights organizations and prompted calls for federal regulation of AI in workplace monitoring.

Root Cause

Amazon's AI productivity tracking system set unrealistic quotas based on top performer metrics without accounting for safety considerations or individual worker capabilities. The system automatically flagged workers for termination without human oversight, creating pressure to work at unsafe speeds to avoid job loss.

Mitigation Analysis

Implementation of mandatory human review for all termination decisions, safety-adjusted productivity targets that account for injury prevention, real-time monitoring of injury rates correlated with productivity metrics, and regular algorithmic audits to ensure quotas remain within safe operational parameters could have prevented harm.

Litigation Outcome

Multiple class action lawsuits filed regarding workplace safety and wrongful termination

Lessons Learned

AI systems making employment decisions require human oversight and safety considerations must be integrated into productivity algorithms. Automated workplace monitoring can create perverse incentives that prioritize metrics over worker safety and wellbeing.

Sources

How Amazon hid its safety crisis
Reveal News · Apr 29, 2019 · news
The Injury Machine: How Amazon's Production System Hurts Workers
Strategic Organizing Center · May 18, 2021 · academic paper