How Cam Stevens Thinks About Voice Technology in EHS

Episode 15 of Headline Podcast featured Cam Stevens, Safety Technologist and CEO of PKG, in a conversation published on March 18, 2026 with Andreza Araujo and Dr. Megan Tranter. Stevens defended a practical thesis that senior EHS leaders should hear before buying another platform: technology helps only when the organization has already named the risk decision behind the technology catalog it wants to improve.

Why voice technology is not a shortcut to safety intelligence

Voice technology in EHS is the use of speech capture, voice assistants, field narration or conversational interfaces to collect safety information from people doing the work. It can reduce friction in reporting, although it can also multiply weak data when leaders treat the tool as the strategy.

On Headline Podcast, Cam Stevens said: "If we're very clear on the problem to solve, then a technology catalog is excellent. If we're not clear on the problem to solve, a technology catalog is very dangerous." That sentence should sit at the front of any EHS technology business case because it separates useful signal capture from vendor-led confusion.

The market often sells voice tools as if the main problem were typing speed. In many plants, mines, warehouses and construction projects, the deeper problem is that workers have learned which risks are worth naming and which risks will be ignored. A microphone cannot fix a leadership system that does not respond.

This is why the discussion belongs beside AI in EHS executive ownership. The technology may be new, but the accountability question is old: who acts when the field says something uncomfortable?

Voice data becomes valuable only after the decision is named

A voice capture pilot should begin with one decision, not a broad promise to modernize reporting. For example, the tool might help classify high-potential near misses, flag repeated permit deviations, capture pre-task concerns before SIMOPS, or identify psychosocial stressors in a large change program.

OSHA describes worker participation as involvement in establishing, operating, evaluating and improving the safety and health program. Voice technology can support that principle when it makes worker knowledge easier to hear, but it weakens the principle when workers speak into a system that produces no visible change.

A senior EHS leader should require 6 fields before approving the pilot: decision, owner, affected work group, data boundary, action threshold and feedback route. If the team cannot complete those fields in plain language, the voice tool is being asked to create governance that leadership has not designed.

The practical question is not whether workers will use the device. The harder question is whether leaders will change work when the device confirms a repeated weak signal.

What the voice channel sees that forms and dashboards miss

Forms usually compress experience into categories, while voice lets people explain context, hesitation, pressure and tradeoffs. A worker might tick "housekeeping" in a digital checklist, but in a spoken note explain that the area is blocked because a contractor stored materials there after a schedule change.

Stevens gave Headline a useful technology test when he said: "We can elevate the human experience with technology, but there are certainly times when technology should absolutely not be used." In EHS, that means voice capture should reduce friction for field truth, not replace a supervisor's obligation to stand where the risk exists.

Voice data can reveal 4 families of signal that dashboards often flatten: repeated language around production pressure, confusion about authority, unclear control status and fatigue or overload in the way work is being described. Those signals connect directly to weak-signal safety dashboard design, where early evidence matters only when it reaches a decision owner.

Do not confuse more words with better intelligence. A thousand short voice notes are noise unless the organization can classify themes, protect privacy, verify field conditions and tell workers what changed because they spoke.

Comparison: voice signal capture versus dashboard theater

Voice technology and conventional dashboards can both improve safety decisions, although each fails in a different way when governance is weak. The useful distinction is whether the tool changes decisions or merely increases the volume of information presented to leaders.

The difference matters because EHS teams already have enough data that nobody owns. Voice technology deserves investment only when it shortens the distance between field evidence and a real control decision.

How leaders should govern privacy, trust and psychosocial risk

Voice records can contain names, health cues, emotional tone, interpersonal conflict, contractor issues and criticism of supervisors. That makes them useful and sensitive at the same time, especially when the tool is introduced during restructuring, staffing pressure or operational change.

NIOSH's 2026 guidance on AI-related workplace hazards treats artificial intelligence as a workplace change that can create new risks if employers do not identify and manage them. A voice-enabled EHS tool should be treated the same way: before launch, assess privacy, work pace, surveillance fear, data accuracy and the effect on worker participation.

Stevens warned that "The changing shift in risk profile will be overwhelmingly psychosocial, driven by technology usage in our organizations." That warning fits the Headline audience because technology can improve safety while also increasing monitoring pressure, response expectations and fear that ordinary speech will be used out of context.

Use a 30-day governance check before the pilot scales. Define what will not be recorded, who can access identifiable audio, when transcripts are anonymized, how long records are kept, which uses are prohibited and how workers can challenge a wrong interpretation. The same logic appears in psychosocial risk from technology audits, where the tool's social effect is part of the risk assessment.

Why field validation still matters after the transcript looks clear

A transcript can sound precise while the underlying condition remains unverified. Voice technology may capture that an interlock was bypassed, a permit step felt rushed or a crew felt short-staffed, but leaders still need field validation before they treat the signal as controlled or closed.

ISO 45001:2018 specifies requirements for an occupational health and safety management system that manages OH&S risks and improves performance. In practice, that means information from a voice channel must enter the same management discipline as any other hazard report: classification, ownership, action, verification and review.

Across 25+ years of executive EHS work, Andreza Araujo has seen that safety information protects people only when it changes routines, authority or resources. Voice notes can expose the gap faster, although they do not close the gap by themselves.

This is where control automation in safety becomes a useful adjacent warning. Automated capture can reduce human effort, but it can also dull human attention if leaders begin to believe the system sees everything that matters.

Recommendation

Senior EHS leaders should pilot voice technology only where the organization can name the decision, protect the speaker and prove that field signals change work. Start with one narrow use case, such as high-potential near-miss narration, pre-task uncertainty during non-routine work or repeated control deviations in a single operating area.

The first 60 days should measure 5 things: number of usable signals, percentage assigned to an owner within 72 hours, percentage verified in the field, number of work changes made and number of workers who received feedback. Adoption alone is a weak success measure because people can use a tool while silently learning that nothing changes.

Do not scale the tool until workers can point to specific examples of risk information that produced a different decision. That is the threshold that turns voice technology from a reporting feature into a safety leadership instrument.

Conclusion

Cam Stevens' Headline argument is useful because it refuses the easy story that better technology automatically means better safety. Voice technology in EHS can make weak signals easier to capture, but only leadership discipline turns those signals into control decisions.

For leaders considering EHS voice tools, the next step is concrete: write the decision rule before the pilot, then test whether the first 60 days produce verified work changes. Listen to the full conversation with Cam Stevens on Headline Podcast and use it to challenge the next technology proposal before it becomes another dashboard.