A few days after cracks appeared along the walls of a Paceville apartment block, it came crushing down. It was a near-miss. No one was killed. But it could easily have gone the other way, and everyone in Malta knows it.

Once again, our country is facing the consequences of a construction sector that is far too comfortable with risk, and a regulatory system too reactive to stop it.

As the blame is passed around — from developers to authorities to outdated laws — one question is gaining urgency: is there anything smarter we could be doing? Could technology help us see the next disaster before it happens?

According to AI expert Prof. Alexiei Dingli, the answer is yes. “AI isn’t a magic wand, but it can serve as an extra pair of eyes, working 24/7, flagging the things humans miss or ignore,” he says.

Dingli has been developing systems for early crack detection in buildings using AI-powered image analysis. But the technology’s potential doesn’t stop there.

Imagine a system that automatically scans every new planning application filed with the Planning Authority. The AI checks for missing documentation, suspicious patterns in excavation plans, or repeat offenders submitting yet another questionable project. It flags the concerns to human regulators before work begins.

“It’s like having a digital watchdog that never sleeps,” Dingli explains. “In a small country like Malta, with limited manpower, that kind of automation could make a real difference.”

The same applies once construction is underway. While site inspections are often infrequent and inconsistent, AI can monitor in real time. A network of cameras, drones and vibration sensors can feed live data to a central system. If a building shifts slightly, or a wall begins to crack, the system can ping authorities instantly. “We already have the tools to track micro-movements in structures,” Dingli says.

“The real challenge is in deciding to use them widely and systematically.”

That challenge is not technical. Most of these systems already exist abroad. Some are even open source or relatively cheap. What’s missing, Dingli argues, is coordination. “It’s not about reinventing the wheel. It’s about stitching together the pieces in a way that works for our specific context. Tight urban zones, old stone buildings, narrow roads, limited regulators.”

AI can also play a role in enforcing safety on construction sites. Computer vision software can scan footage to check if workers are wearing helmets and harnesses, whether trucks are speeding, or if barriers are in place. In unsafe conditions, a real-time alert can be sent directly to the site manager. It’s a form of constant oversight that doesn’t rely on chance inspections.

Crucially, the data doesn’t have to come only from machines. Malta’s residents are already reporting warning signs. On social media, in planning objections, and on helplines. AI can help parse these streams of information and spot patterns. If ten complaints about shaking walls appear across one neighbourhood in two days, the system can flag a possible emerging threat. “People’s observations are often dismissed,” Dingli says. “But when clustered together and analysed properly, they become a valuable early-warning system.”