By the Numbers
A Closer Look at the Data
The AI-EHS Adoption Journey
Share of organizations at each stage of AI adoption for EHS
Bar chart of AI adoption stages in EHS: 12% not using or exploring, 28% discovery, 30% piloting, 23% scaling and 5% integrated.
58% Exploratory
29% Transformative
12%
28%
30%
23%
5%
Not currently using or exploring
Discovery/ understanding
Piloting/ experimenting
Scaling
Integrated
Adoption phase
Most organizations are in the exploratory stages of AI adoption. Not sure = 1%.
View data table
| Adoption stage | Share of organizations |
|---|
| Not currently using or exploring | 12% |
| Discovery/understanding | 28% |
| Piloting/experimenting | 30% |
| Scaling | 23% |
| Integrated | 5% |
| Exploratory (discovery + piloting) | 58% |
| Transformative (scaling + integrated) | 29% |
Goals vs. Outcomes
What organizations hoped AI would do for EHS, compared with what they report it has done
Chart comparing AI-EHS goals with realized outcomes. Reducing repetitive work and gaining insights meet or exceed goals at every stage. Lowering SIF rates is still developing: a 33% goal vs. 12% outcome at scaling and 26% at integration.
Outcomes: scaling
Outcomes: integrated
Goals (discovery/piloting)
Reduce time spent on repetitive tasks
Get more actionable insights from data
Stay up to date on EHS standards/regulations
Predict risk with greater accuracy
Assess or respond to risks in real time
Time savings and data insights are goals achieved early in the adoption cycle. More ambitious outcomes, like predicting risk with greater accuracy and responding to risks in real time, are generally not achieved until late stage integration.
View data table
| Goals (discovery/ piloting) | Outcomes (scaling) | Outcomes (integrated) |
|---|
| Reduce time spent on repetitive tasks | 75% | 78% | 79% |
| Get more actionable insights from data | 64% | 68% | 77% |
| Stay up to date on EHS standards/regulations | 51% | 38% | 57% |
| Predict risk with greater accuracy | 50% | 35% | 51% |
| Assess or respond to risks in real time | 31% | 30% | 58% |
| Lower our costs | 33% | 24% | 38% |
| Lower SIF incident rates | 33% | 12% | 26% |
Top 5 Barriers to AI Adoption
Share of organizations citing each barrier
Horizontal bar chart of the Top 5 barriers to AI adoption in EHS, led by low resource investment at 74% and data quality at 62%.
Low/nascent resource investment
Lack of ownership/champion/priority
Mistrust/discomfort/lack of confidence
Concerns about data privacy, security or bias
Resources top the list of barriers, and data quality emerges as a significant challenge. Both are areas where organizations can start preparing today.
View data table
| Barrier | Share of organizations |
|---|
| Low/nascent resource investment | 74% |
| Data quality | 62% |
| Lack of ownership/champion/priority | 59% |
| Mistrust/discomfort/lack of confidence | 55% |
| Concerns about data privacy, security or bias | 55% |
The Training Effect
Safety professionals with and without AI training
Comparison showing safety professionals with AI training report higher AI knowledge (75% vs. 56%), comfort interpreting AI metrics (79% vs. 61%) and belief that AI improves safety decisions (76% vs. 59%).
No AI training
AI training
AI-EHS knowledge
(very/somewhat)
56%75%
Comfort interpreting AI metrics
(very/somewhat)
61%79%
AI improves safety decision-making
(significantly/somewhat)
59%76%
0%25%50%75%100%
Safety professionals with AI training rate themselves higher on every AI-readiness measure.
View data table
| Measure | No AI training | AI training |
|---|
| AI-EHS knowledge (very/somewhat) | 56% | 75% |
| Comfort interpreting AI metrics (very/somewhat) | 61% | 79% |
| AI improves safety decision-making (significantly/somewhat) | 59% | 76% |