Day 1: Building the Safety Prediction Evidence Base

Hazard and Incident Taxonomy

• Performance signals around incident taxonomy
• Leading indicators of unsafe behavior
• Preventive actions ranked by effectiveness
• Worker exposure verified before conclusions

Exposure and Leading Indicators

• Practical tests for leading indicators
• Incident patterns revealing control weakness
• Investigation evidence preserving causal context
• Control failures triggering timely escalation

Safety Data Quality

• Practical tests for safety data quality
• Hazard pathways and exposure conditions
• Escalation thresholds for safety alerts
• Risk controls matched to hazard severity

Day 2: Identifying Elevated Safety Risk

Predictive Risk Pattern Detection

• Ownership expectations for predictive risk
• Residual risks recorded after mitigation
• Evidence required for risk assessment
• Human review of critical safety alerts

Near-Miss and Condition Signals

• Validation checks for condition signals
• Corrective actions assigned to accountable owners
• Control effectiveness before residual risk
• Near-miss evidence informing prevention

Safety Risk Thresholds and Alerts

• Performance signals around safety risk thresholds
• Safety observations checked against field conditions
• Accountability for approved AI outputs
• Governance roles across the lifecycle

Day 3: Converting Predictions into Prevention

Preventive Action Prioritization

• Ownership expectations for preventive action prioritization
• Policy limits limiting unauthorized use
• Independent review for high-impact cases
• Human oversight at high-impact decisions

Supervisor Review and Escalation

• Ownership expectations for supervisor review
• Exceptions handled with recorded rationale
• Escalation routes for control failures
• Audit evidence supporting full traceability

Predictive Safety Control Plan

• Application limits around predictive safety
• Oversight cadence for governance controls
• Decision rights and approval thresholds
• Control testing prior to workplace approval

• Case analysis examining leading indicators in workplace decisions
• Simulation testing condition signals under operational pressure
• Guided exercise applying preventive action prioritization to workplace cases
• Group review evaluating predictive safety against evidence