AI Is Now Routine in Safety Programs, but Control Questions Persist
Artificial intelligence has moved from pilot projects into core safety operations across most organizations, according to a 2026 survey by Wolters Kluwer and the National Safety Council. Eighty-two percent of EHS leaders report at least moderate AI use, with one in five saying they've achieved extensive integration. Only 2% have no AI activity.
The adoption rate is accelerating. Ninety percent of respondents plan moderate to significant AI investment in the next period.
But speed is outpacing confidence. Ninety percent of respondents flagged at least one concern about AI in their EHS programs, even among those who consider themselves fully prepared.
Incident Prevention Leads, but Trust Gaps Remain
The top stated benefit of AI in EHS is its ability to predict and prevent incidents, cited by 30% of respondents. Improved efficiency in reporting and compliance came second at 26%. Organizations are deploying AI in risk assessment, hazard identification, data analytics, and predictive maintenance scheduling.
The leading worry is overreliance on AI at the expense of human judgment. Fifty-one percent of respondents cited this concern. Data privacy and security risks followed at 50%, with poor-quality data flagged by 42%.
The pattern is clear: EHS professionals want AI to inform decisions, not make them. The survey recommended that leaders explicitly define where AI serves as decision support and where human accountability must remain.
Digital Gaps Undermine AI's Value
Organizations are unevenly digitalized, limiting what AI can actually do. Only 11% have fully digital, integrated systems. Thirty-seven percent mix digital and manual processes, while 18% still rely heavily on paper.
Foundational compliance work is furthest along: 58% have digitalized safety data sheet management, 54% have digital inspections and audits, and 54% have digital regulatory reporting.
But the processes closest to daily operations lag far behind. Only 36% have digitalized behavior-based safety observations, 35% track occupational health monitoring digitally, and 30% have digital permit-to-work systems. These behavioral and operational areas are the ones most likely to surface early risk signals.
Mental Health Acknowledged but Deprioritized
EHS mandates are expanding to include psychosocial safety, mental health, fatigue, and hybrid-work risks. Eighty-seven percent of respondents agreed that mental health falls within the safety function's scope, with 62% strongly agreeing.
Yet when asked to rank challenges, respondents placed infectious disease preparedness and aging workforce concerns above mental health and workplace stress. The gap between acknowledgment and action reflects what the survey called "a crowded agenda where human-centric risks are acknowledged yet often deferred."
Skills Shift Signals AI's Integration
The profession's skill requirements are changing fast. When asked what capabilities junior EHS professionals need most, respondents ranked understanding AI use at 48%, followed by digital literacy at 45% and data analysis at 42%. These ranked above traditional competencies like regulatory expertise and field experience.
Nearly half of respondents said their organizations plan to use AI agents and automation to offset the impact of retiring personnel.
For insurance professionals, the findings point to a sector in transition. AI is embedded in safety operations, but governance frameworks haven't caught up. The risk of poor decisions based on incomplete data or unchecked algorithmic recommendations is real - and it's a concern that spans the industry.
Your membership also unlocks:
