This blog post is the final segment in a three-part series highlighting key takeaways from Merlin’s recently published whitepaper, “Innovate to Elevate.”

Effective communication is the backbone of aviation safety. But it isn’t foolproof: miscommunications over the radio remain a persistent challenge in the National Airspace, and in recent years, they’ve been implicated in a troubling rise in runway incursions.

Between January and October 2023, the FAA recorded 19 serious runway incursions, the most in any 10-month period since 2016. Many of these incidents involved missed or misunderstood radio calls. In response, the FAA launched an independent review and, in 2024, opened an audit of 45 of the country’s busiest airports. The National Transportation Safety Board (NTSB) has since echoed these concerns, noting that while serious near-miss events are rare, the aviation safety system is “showing clear signs of strain.”

It’s time to rethink how we support communication in the cockpit and control tower.

At Merlin, we believe advanced automation tools can play a critical role. We’re developing a communication assistance system that uses natural language processing (NLP) to interpret real-time radio transmissions and offer intelligent, context-aware recommendations to pilots. These systems help pilots understand when a call is relevant to their flight, process communications from nearby aircraft, and build a clearer operational picture.

An NLP-based assistant could alert pilots if, for example, they cross a runway hold short line without a proper clearance, or attempt a landing without authorization. These tools can also enhance situational awareness at non-towered airports, where visibility of non-cooperative traffic can be limited.

As we continue integrating increasingly automated systems into the National Airspace, improving how humans and machines communicate will be just as important as improving how aircraft fly. The next era of aviation safety will depend not just on better flying, but also on clearer listening.