'We can’t have eyes everywhere 24-7, but the AI can help,' says company's digital expert
EllisDon, a major Canadian construction firm, integrates artificial intelligence into its health and safety protocols, aiming to set a new industry benchmark for responsible technology use. The company deploys computer vision systems that analyze images and video from both interior and exterior cameras on active job sites. These systems detect hazards such as missing guardrails, misplaced materials, and workers not wearing required safety gear, allowing supervisors to address risks before they escalate.
AI-powered hazard detection and monitoring
Patrick LaLonde, senior director of digital project delivery, describes the process: “We do 360 photo walkthroughs. Essentially, we’ll take a camera, walk through our job site on a regular basis with a 360 photo, and it kind of creates like this Google Street View sort of version of the job site. When we upload that to our system, it’s able to analyze each photo and pick out safety infractions.”
EllisDon’s technology extends beyond hazard detection. The AI tracks deliveries, counts personnel and vehicles, and compiles statistics on site activity. This data enables remote monitoring of hard-to-reach locations, providing management with real-time insights into operations. The system also stores historical footage, allowing safety teams to review incidents and identify patterns that could inform future prevention strategies.
The company’s approach to AI is deliberately supportive rather than substitutive. LaLonde says, “Our approach will always be an assistance to our traditional process to help us be more accurate and be more efficient with our reporting and creating the safer job sites. But not in any way a replacement to any human interaction.”
Prioritizing privacy and responsible governance
Worker privacy remains a priority. EllisDon relies on traditional site cameras, which have long been standard in construction, and avoids more invasive technologies that require personal information. “We did see some resistance from tools like [workforce management systems] because we had people not willing to put their cell phone number in there to create their account and not wanting to share email addresses or even information like that. So it was just a bit more intrusive, a bit more personal,” LaLonde explains.
To manage the evolving landscape of AI applications, EllisDon develops governance policies that address each new use case individually. The company reviews potential implementations for ethical risks and data sensitivity, ensuring that every deployment aligns with its standards for transparency and accountability. LaLonde notes, “We do have like a governance policy that’s sort of in development. The AI use cases are continuing to grow and evolve. We’re absolutely taking those case by case and sort of reviewing those and treating that with caution.”
Trying to set an example
EllisDon also observes broader industry trends, as construction firms increasingly use AI to automate repetitive tasks and streamline communication between teams. The company believes that AI’s greatest value lies in applications that require constant vigilance, such as site monitoring, where machines can operate continuously and consistently.
LaLonde encourages safety leaders to focus on targeted, practical uses for AI. “The best application right now is reducing a lot of that manual human reliance on things like site monitoring and repetitive tasks. We can’t have eyes everywhere 24-7. A human can’t, but the AI can help there. It can help bridge that gap,” he says.
By embedding AI into its safety culture, EllisDon demonstrates that technology and ethics can advance together. The company’s transparent, case-by-case approach to AI adoption offers a model for responsible innovation in construction and beyond.
