What's Inside

This white paper examines real failures from live U.S. patrol robot deployments, including:

• A case where robots kept operating—but operators lost control
• AI systems that appeared normal—yet made unsafe or unexplainable decisions
• How one privacy complaint can immediately pause patrol operations
• Documented incidents where robots became unintended surveillance devices

This Paper Helps Operators

• Identify where control is lost even when robots remain operational
• Prevent privacy issues from escalating into immediate patrol shutdowns
• Recognize unsafe or drifting AI behavior during unsupervised hours
• Anticipate real-world risks that never appear in pilot programs