The Future of Workplace Safety: How AI is Changing the Game

In today’s complex work environments, technology is redefining how we approach safety. To mark the World Day for Safety at Work on April 28, 2025, we joined Bonava—one of Europe’s leading residential developers—for a webinar where our CEO and founder, Lamin Faye, shared how AI is transforming workplace safety.

Why Workplace Safety Needs a Rethink

Workplace accidents cost around 4% of global GDP annually. Despite decades of training and procedures, incident rates in Europe have plateaued over the past 15 years.

The message is clear: traditional safety measures are no longer enough.

Enter AI, And more specifically, Computer Vision—a proven AI technology that enables machines to interpret visual information like humans, but with greater accuracy and consistency.

Computer Vision works because its always on, meaning AI systems don't fatigue in the way humans do, they can monitor worksites 24/7 with exceedingly high accuracy. Computer Vision now outperforms humans in object detection with 95% accuracy. And it can identify patterns like recurring risks in a way that manual inspections often miss.

The Competitive Advantage of AI-Powered Safety

With this greater accuracy and consistency, the benefits of AI go beyond just a compliance perspective—for organisations they deliver real operational value such as:

• Fewer Accidents: Leading to reduced downtime and fewer claims
• Improved Compliance: Staying on top of regulation
• Better Decisions: Data reveals where resources are most needed
• Reputation Boost: Demonstrates commitment to employee wellbeing
• Cultural Shift: Moves organizations from reactive to proactive safety

From Insight to Action: AI in the field

Computer vision is at the core of our platform—and it’s already delivering tangible safety improvements across real worksites helping customers make smarter, faster, safety decisions every day:

• PPE Detection in Real Time
  Instantly spots when workers are missing helmets, vests, or gloves—allowing HSE personnel to intervene before risk escalates.
• Clear Emergency Exits
  Flags blocked exits so they can be cleared before they become a safety hazard.
• Vehicle and Pedestrian Safety
  Monitors proximity between people and moving vehicles like forklifts or trucks, alerting teams when either gets too close.
• Smarter Site Layouts
  AI has prompted changes like relocating machine on/off buttons that were being accessed in unsafe zones, moving containers out of pedestrian paths, and creating dedicated walkways to minimize vehicle interaction.
• Improved Signage and Markings
  Customers have updated floor markings and added safety signage in high-risk areas based on heatmaps and incident data.

These seemingly small changes—guided by data—are reducing risks, improving compliance, and helping build a stronger safety culture on site.

Tackling the ‘Big Brother’ Concerns

But as with any new technology, concerns exist and trust matters. In our industry the concerns raised are mainly around privacy and surveillance. When talking to companies exploring AI implementations in workplace safety we typically hear the following concerns: How is data collected? Are employees being monitored? How does this align with GDPR?

These questions are valid and we address them head-on and build trust in AI through:

1. Data Anonymization

Unlike traditional security cameras, modern safety systems can protect identity while detecting risks. "We don't have a live feed. We blur everything in the image, and we constantly innovate better ways to anonymize," Lamin noted in the webinar. This fundamental shift from identifying individuals to identifying situations represents a shift in workplace monitoring.

2. Data Minimization

The responsible approach follows a simple rule: only collect what you need. We don't store any data unless there is a flagged risk, a situation which requires an observation. This minimizes unnecessary data collection while still capturing valuable safety insights.

3. People-Centered Implementation

Technology deployment should never happen in isolation from the people it affects. Early engagement with stakeholders—including unions and workers' councils—builds understanding and trust. We recommend demonstrating the technology, explaining its limitations, and being transparent about how the data collected will be used.

The distinction between surveillance and safety monitoring is crucial. This is a different kind of use case. We're using cameras as sensors. Not as a surveillance tool. By keeping people at the center of technological implementation, organizations can harness AI while respecting worker privacy and maintaining GDPR compliance.

From Detection to Prevention: The New Safety Stack

Moving from concerns to future possibilities, the real power of AI lies in its ability to go beyond detection and into prevention. This shift happens through what we believe are four key stages:

1. Data Collection – Vision systems capture visual safety data continuously
2. Analysis – Dashboards and heat maps reveal risk patterns
3. Recommendations – AI suggests targeted interventions
4. Automation – Future systems will even act independently to reduce risks

“We believe that the AI-enabled EHS system will become a standard. We see AI acting as a Safety Co-pilot—guiding teams on the ground and helping EHS managers make better, faster decisions. For those concerned that AI will take your job. It won't. But someone using AI will." Lamin explained.

The Future of AI in Workplace Safety

Looking ahead, new technologies will continue enhancing how the industry as a whole will protect workers.

Smart wearables will evolve beyond simple data collection tools to become dynamic communication devices that can interact with workers in real time—alerting them to hazards or guiding them through safe procedures. Environmental sensors, when integrated with computer vision systems, will enhance the ability to detect dangerous conditions such as gas leaks or excessive heat, providing a more comprehensive safety net. Augmented reality (AR) and virtual reality (VR) will revolutionize training by creating immersive, realistic environments where workers can safely practice handling high-risk scenarios or receive real-time guidance during complex tasks.

Finally, predictive analytics will play an increasingly important role, by analyzing historical safety data to identify patterns and forecast potential risks—enabling organizations to take preventive action before incidents occur.

Conclusion

The era of reactive safety management is ending. With mature AI technologies like computer vision now readily available, organizations have powerful tools to prevent accidents before they happen.

As we observed during World Day for Safety at Work, the question is no longer whether AI has a place in workplace safety—it's how quickly organizations can implement it. Making a start can seem daunting at first, but with the right partner we know it’s possible. Starting small, engaging stakeholders, focusing on using AI insights to drive continuous improvement ad building trust are all steps in the right direction.

The future of workplace safety is proactive, data-driven, and powered by AI.