C4: From Insight to Intervention With AI Safety
Mon, 6/1: 2:00 PM - 3:00 PM CDT
Education Sessions
Education Sessions
Ernest N. Morial New Orleans Convention Center
Room: 293
CM Credit Hours: 1
Description
Despite decades of progress, serious injuries and fatalities (SIFs) remain stubborn. The barrier is no longer only controls, it is knowing which low frequency, high consequence risks to prioritize and how to stop them in motion. This session will show how organizations are doing both, using AI. This updated program goes beyond concepts to field results.
We will:
Extract and rank SIF precursors from incident narratives with Natural Language Processing (NLP) and large language models.
Prioritize risks with robust, transparent statistical methods, including Bayesian hierarchical models that blend company and industry data to overcome bias and scarcity.
Prevent collisions in real operations using computer vision and 3D sensing (monocular depth, stereo, LiDAR), with live/anonymized clips of detections, smart danger zones, and alarm strategies tuned for minimal false alarms.
Participants will see cross industry examples and a practical playbook for piloting, scaling, and sustaining AI safety systems. They will leave with tools to identify underappreciated risks, prove priorities, and intervene before harm.
We will:
Extract and rank SIF precursors from incident narratives with Natural Language Processing (NLP) and large language models.
Prioritize risks with robust, transparent statistical methods, including Bayesian hierarchical models that blend company and industry data to overcome bias and scarcity.
Prevent collisions in real operations using computer vision and 3D sensing (monocular depth, stereo, LiDAR), with live/anonymized clips of detections, smart danger zones, and alarm strategies tuned for minimal false alarms.
Participants will see cross industry examples and a practical playbook for piloting, scaling, and sustaining AI safety systems. They will leave with tools to identify underappreciated risks, prove priorities, and intervene before harm.
Learning Outcomes
Upon completion, the participant will be able to:
• Apply computer vision to detect high risk people/vehicle interactions and configure smart zones that adapt to speed and heading.
• Analyze using AI to mine data for SIF precursors Prioritize risks with Bayesian models that combine local and industry evidence.
• Explain results to leaders and frontline teams.
• Recognize AI monitoring, including metrics (precision/recall, false alarm rate, intervention effectiveness) and change management steps.
• Mitigate cognitive biases (e.g., availability bias) with data driven ranking and transparent model outputs.
• Produce task specific safety briefings that connect ranked risks to practical controls, training, and checklists.
• Apply computer vision to detect high risk people/vehicle interactions and configure smart zones that adapt to speed and heading.
• Analyze using AI to mine data for SIF precursors Prioritize risks with Bayesian models that combine local and industry evidence.
• Explain results to leaders and frontline teams.
• Recognize AI monitoring, including metrics (precision/recall, false alarm rate, intervention effectiveness) and change management steps.
• Mitigate cognitive biases (e.g., availability bias) with data driven ranking and transparent model outputs.
• Produce task specific safety briefings that connect ranked risks to practical controls, training, and checklists.
Content Level
Intermediate
Core Competencies
Risk Assessment
Risk Management
Keywords
Respiratory protection
Risk assessment and management
Safety
Session Availability
In-person
OnDemand
Virtual
Targeted Audience
Professional
Transfer of Knowledge
Case Studies
Practical Application