AI is coming to construction sites. But does it make workers safer?

Key points

• FYLD, a London-based software company, uses AI to help construction and utility workers predict permits, plan around weather and cut paperwork.
• Amey will roll FYLD out to 2,500 field workers, and Copsey said case studies have reported reductions in incidents and injuries of as much as 48%.
• Researchers say evidence that AI actually prevents accidents is still limited and warn against relying on it in place of human supervision.

Artificial intelligence is increasingly being used on construction sites and in other physically demanding jobs, with companies saying the technology can cut paperwork and reduce accidents. FYLD, a London-headquartered software company led by CEO Shelley Copsey, uses AI to support front-line workers in construction, utilities and infrastructure.

Copsey told CNBC that attention usually turns to budgets or cash flow when fieldwork goes wrong. She said the focus should instead be on what happens in front-line workers' daily work and how to make their jobs easier.

How the tools work

FYLD uses AI to predict which permits a project will need, helping workers avoid delays. Copsey said workers previously often arrived at a site without all the information they needed. The system can also help crews plan around the weather, such as moving heavy lifting away from a day when strong winds are forecast.

Workers record short videos describing their tasks and safety concerns, and AI analyzes the information to help workers and managers anticipate potential problems. Copsey said field workers can spend around 20% to 25% of their time on site on administrative work, and the system can generate draft reports, reducing a task that took around half an hour to two or three minutes.

Treat AI as an extra pair of eyes, not a replacement for supervision

— Waleed Umer

Doubts over safety gains

Infrastructure services provider Amey announced in March that it would roll FYLD out to 2,500 field workers. Copsey said FYLD case studies have reported reductions in incidents and injuries of as much as 48%. AI is also used to reduce physical inspections: Bentley Systems' Blyncsy platform analyzes images captured by drivers to spot problems such as damaged guardrails and debris.

Waleed Umer, an assistant professor at Northumbria University, told CNBC that many studies have focused on what models can detect rather than whether they prevent harm. Billy Hare, a professor at Glasgow Caledonian University, said research has mainly focused on productivity gains. The U.K. Health and Safety Executive recorded 25 construction worker deaths in 2025/26, down from 35 the previous year, but Hare said an economic slowdown may explain the fall.

Calls for human oversight

Umer said human oversight may not be enough if AI produces more information than supervisors can review. He said AI should be treated as an extra pair of eyes, not a replacement for supervision. Ultimo CEO Steven Elsham said companies should decide what time saved is for and measure safety and first-time fixes rather than just output.