The construction industry has always known that safety depends on trust. A worker who doesn’t believe the system is on their side often won’t report a near miss, raise a hazard, or engage with a safety process that feels like surveillance. Now, as AI tools move onto job sites, that dynamic is being tested in new ways. Recent controversies surrounding AI-powered employee monitoring programs have exposed a reality many businesses are in the early stages of confronting, raising questions around surveillance, privacy, and worker autonomy. The technology can do remarkable things. But it only works if the people it’s designed to protect are willing to use it.
This challenge runs especially deep in construction and other hazardous industries, where AI is being introduced to identify safety threats, surface emerging risks, and improve incident prevention. While the technology is advancing, the adoption is not always keeping pace. That gap often comes down to frontline workers who read these tools as “Big Brother” surveillance of their behavior, rather than a resource designed to protect them. How teams navigate that perception will shape whether AI meaningfully advances safety outcomes or becomes another layer of workforce resistance and friction on an already demanding job.
The same technology can feel like protection or surveillance
Most workplace AI systems rely on some form of data collection. That data may come from safety observations, incident reports, environmental sensors, cameras, wearable devices, or digital workflows. More information should, in principle, help organizations identify patterns that lead to injuries and accidents.
The problem is that frontline workers experience data collection in practice, not in theory.
When employees believe technology is being used primarily to measure productivity, track movement, or evaluate performance, trust erodes quickly. Even systems designed with safety objectives can trigger skepticism if workers are unclear about what information is being collected, who has access to it, and how it will ultimately be used.
The Edelman Trust Barometer consistently shows that trust in employers remains higher than trust in most institutions, but that trust is fragile when organizations fail to communicate transparently about new technologies. Once workers begin to see safety tools as a means of surveillance, adoption gets much harder. Worker behavior directly influences the effectiveness of any safety program.
Safety depends on participation
Many serious workplace incidents tend to be preceded by warning signs that never make it into formal reporting systems. A worker might notice a recurring hazard but never report it. Sometimes a supervisor sees a near miss and has no easy way to document it, or a crew quietly develops a workaround for a recurring problem that never travels beyond their immediate team. These information gaps exist in nearly every industry, from manufacturing and logistics to construction and energy. The value of AI in safety is its ability to surface patterns across thousands of observations that human reviewers may never connect manually. However, those insights only exist if workers are willing to contribute information in the first place.
If reporting systems become associated with monitoring or discipline, participation declines. Workers are less likely to share observations, document concerns, and engage with the process. The result is a weaker dataset and less visibility into emerging risks. Organizations often focus on what AI can analyze while overlooking the more fundamental question of whether workers are willing to provide meaningful information to analyze at all.
Visibility and trust are not opposing goals
Effective safety programs have always required both operational visibility and worker trust, and AI makes that more visible than ever. Safety leaders need visibility into conditions across large facilities, multiple shifts, and geographically distributed operations. Manual inspections and periodic audits can only capture a fraction of what occurs during a typical workday. AI offers the potential to identify trends and hazards at a scale that traditional processes struggle to match.
But visibility becomes problematic when workers feel they’re the primary subject of observation rather than the environment around them. The distinction matters. A system designed to understand workplace risks focuseson conditions, hazards, trends, and prevention opportunities. One built primarily to evaluate employee behavior sends a very different message.
The most successful safety technologies are often those that make workers feel empowered. They help employees communicate concerns, document risks, and contribute expertise from the field. The technology becomesa mechanism for participation. That shift in framing can dramatically influence adoption.
Transparency is becoming a safety requirement
Businesses introducing AI into safety workflows often concentrate on implementation details such as integrations, reporting capabilities, and deployment timelines. Many spend considerably less time explaining the system to the workforce that will interact with it daily. That communication gap creates unnecessary risk.
Workers increasingly want answers to straightforward questions about the data being collected, who can see it, if it can be used against them, and how it actually makes their job safer. When those questions go unanswered, employees fill in the blanks themselves. And they rarely fill them in favorably.
Transparency doesn’t require organizations to disclose every technical detail of an AI model. It requires clarity around purpose, governance, and accountability. The National Institute for Standards and Technology’s AI Risk Management Framework emphasizes transparency and explainability as foundational principles for responsible AI deployment. These principles become especially important when systems influence workplace decision-making. Workers are more likely to support technology when they understand its intended function and can see clear safeguards around its use.
The future of safety AI is collaborative
Many discussions around workplace AI focus on automation replacing human judgment, but safety presents a different opportunity. The most effective safety programs have always relied on frontline expertise. Workers closest to the job often identify emerging risks long before those risks appear in reports or dashboards. AI can strengthen that expertise rather than replace it.
For example, a worker may submit a safety observation through a mobile device or voice report. AI can help categorize the information, identify similar reports across multiple locations, and surface patterns that would otherwise remain hidden. Safety leaders gain broader visibility while workers retain ownership of the knowledge they contribute. Technology, in this model, acts as an amplifier for human insight.
That distinction is critical because safety culture is ultimately built through engagement; automation alone won’t create it. Organizations with strong safety cultures create environments where workers feel comfortable speaking up, reporting concerns, and participating in risk reduction efforts. AI can support those objectives, but it can’t substitute for them.
Why engagement outperforms monitoring
The long-term effectiveness of workplace safety programs depends more on engagement than on enforcement. Monitoring surfaces what happened; understanding why it happened takes engagement.
A camera may detect a worker entering a restricted area, but it can’t reveal that production pressures, staffing shortages, or unclear procedures drove the decision. Fatigue indicators picked up by a wearable device tell a similar story: the tool can’t address the scheduling practices or operational conditions that create the fatigue in the first place.
Incident reduction requires an understanding of that context. AI earns its place when it helps teams see that context, and that happens only when workers actively participate in the process. That participation breaks down when employees feel technology has been deployed on them rather than with them. This is where many workplace AI initiatives run into trouble. Leaders invest in the tools and underestimate the people-side of making them work.
The organizations achieving the greatest value from safety technology are often those investing as heavily in trust-building as they are in software deployment.
Trust is becoming a competitive advantage
AI adoption is accelerating across industries, and the organizations that succeed will be those that create environments where workers are willing to share meaningful information.
Better participation will lead to stronger reporting, better insights, and a more effective prevention strategy. The result, over time, is a safer workplace and a more resilient safety culture. The alternative is a cycle where employees disengage, data quality declines, and technology produces diminishing returns despite increasing investment.
Workplace safety has always depended on trust between organizations and their employees. AI doesn’t change that reality; it makes it more important.
The next phase of workplace safety technology will be defined by whether organizations can deploy these tools in ways that workers view as supportive, transparent, and trustworthy. That comes down to how leadership presents and manages these tools. The companies that get this right will see stronger participation, better data, and safer outcomes. Those who ignore it may discover that the greatest obstacle to AI adoption was never technical capability. It was human trust.
Author Bio
Cory Linton is a seasoned software industry leader with nearly three decades of experience, including roles at Microsoft and School Improvement Network. In 2018, he co-founded Safety Mojo to empower safety professionals with innovative, data-driven solutions. He has extensive expertise in strategy execution, sales, marketing and communications, and product development. Cory received his MBA from Columbia University and has an undergraduate degree in Latin and Roman History. Cory resides in Salt Lake City with his wife and five children.
