Air traffic agencies test AI help as flights grow, trust remains a hurdle

In short: Air traffic control is looking at AI tools to handle more flights, but safety and trust concerns mean humans are still expected to lead.

What's going on

Air travel has grown since the Covid-19 pandemic, and experts expect more flights in the coming years. At the same time, air traffic control is short on staff, and training a controller can take years and cost hundreds of thousands of pounds, according to the Financial Times report.

Several groups are exploring whether AI can help controllers do their jobs. The idea is that computers could process large amounts of flight, weather, and route data and warn earlier about “conflicts,” meaning situations that could lead to a collision. Think of it like a very fast assistant that highlights possible problems before a human would notice them.

In the UK, National Air Traffic Services (NATS) is running Project Bluebird to study this. NATS says it will use “digital twinning” and “machine learning.” A digital twin is like a realistic simulator of air traffic built from past data, and machine learning is a way for software to learn patterns from that data. NATS says the goal is to advise controllers, not replace them.

Research suggests trust is still a major issue. A study by Chinese academics found controllers were wary of AI suggestions, and that simply shifting humans into a role of checking AI advice might not save much work.

What to watch

Even supporters say it could take years of testing and approval to prove these systems are safe. Another concern is backup plans, since aviation depends on having fallbacks when systems fail, like keeping older navigation tools if GPS is jammed or unavailable.

Source: Financial Times