Few Safety Programs Fully Integrate A.I.
Only 5% of organizations have fully integrated A.I. into environmental health and safety, with most still in discovery or piloting.
Updated on Nov. 5, 2026 in Artificial Intelligence
Live Poll
Would you trust AI to make workplace safety decisions about you without human review?
The American Society of Safety Professionals released a survey on artificial intelligence adoption in environmental health and safety (EHS) programs, drawing on nearly 1,000 safety professionals surveyed online from July 6 through Aug. 2, 2026. Only 5% of organizations consider their EHS programs fully integrated with A.I.; the technology is research-stage in most of these programs, with 58% still in discovery or piloting.
Why it matters
The survey maps a gap between adoption and results: cost savings and reductions in serious injury and fatality rates remain largely out of reach, which ASSP partly attributes to upfront investment costs outpacing returns. Because two-thirds of organizations began their A.I.-EHS work within the last 18 months, the field is still early in its trajectory.
The survey measured organizational maturity across five stages of A.I.-EHS adoption: non-adopters, discovery, piloting, scaling and integrated. Only 5% reach full integration, 58% remain in discovery or piloting, and the stage-by-size gap is stark, with 45% of organizations over 5,000 employees classified as transformative against 19% of those under 250. Whether the projected rise to 1 in 5 integrated organizations within two years materializes has not yet been tested.
The players
American Society of Safety Professionals (ASSP)
A professional association for safety practitioners whose survey defines the five-stage framework used to classify A.I. adoption in EHS programs.
The details
ASSP sorted organizations into five stages of A.I.-EHS adoption, where EHS refers to environmental health and safety programs that manage workplace hazards and compliance. Organizations at the discovery stage are still evaluating the technology, while piloting organizations test it in limited settings before scaling it across operations. The reported gap between stages is largely a question of data and governance: only 33% of discovery-stage organizations rate their data quality as good or excellent, and 62% of integrated organizations have adopted A.I. ethics policies compared with 20% at the discovery stage.
Timeline
Online survey conducted from July 6 through Aug. 2, 2026.
Two-thirds of organizations began their A.I.-EHS work within the last 18 months.
More than half of organizations expect to reach scaling or integrated stages within two years.
The Tech Race
The ASSP framework gives the field a common vocabulary for measuring A.I. maturity in workplace safety rather than a single benchmark of product performance. The results show that adoption is concentrated in large organizations and insurers, leaving smaller firms behind in the transition to scaled use.
Workers in large organizations and insurance firms are the most likely to see A.I. tools in their safety programs first, since those groups lead the transformative classification. Smaller employers should expect a slower timeline, as only 19% of organizations under 250 employees have reached transformative status.
The takeaway
Organizations expect to move from pilot to integrated A.I. programs within one to two years, so watch whether the projected rise from 1 in 20 to 1 in 5 fully integrated organizations occurs in ASSP's follow-up findings. Track whether data-quality ratings improve among discovery-stage organizations, since that is where the survey identifies the weakest foundation.
Further reading
For broader context on how these tools are being deployed, see the latest Artificial Intelligence coverage.
Source note: This article includes information reported by Risk & Insurance.
Live Poll
Would you trust AI to make workplace safety decisions about you without human review?
