AI Arrives at Construction Sites. But Does It Make Workers Safer?
Companies are using AI to plan work, reduce paperwork, and detect potential threats at construction sites and infrastructure facilities. Researchers warn that there is still insufficient data confirming a reduction in accidents due to such systems.
FYLD uses AI to support workers in construction, utilities, and infrastructure projects. The system predicts required permits, helps account for weather conditions, and analyzes short videos in which employees describe tasks and safety issues. According to FYLD CEO Shelly Copsey, photos, videos, and work data reduce report preparation time from about 30 minutes to two or three minutes.
In March, infrastructure services provider Amey announced the implementation of FYLD for 2,500 field employees. Copsey stated that in certain cases, companies reported a reduction in incidents and injuries of up to 48%.
Other developers use AI to inspect sites and identify missed incidents. Bentley Systems' Blyncsy platform analyzes images taken by drivers to detect damaged fencing, overgrown vegetation, and litter. Ultimo tools review work assignments and shift logs, after which they can draft an incident report for human review.
Researchers note that such solutions do not guarantee risk reduction. According to Northumbria University assistant professor Walid Umer, cameras do not see every hazard in a constantly changing construction site, and a high number of false alerts may lead to ignoring signals. He advises treating AI as an additional control tool rather than a replacement for supervision.
