# AI Finds 15,000 Hidden Near-Misses — FAA Rewrites the Rules

- **Date:** 2026-03-19
- **Author:** Samuel Chandra (Founder & CEO, Deepsky)
- **Canonical:** [https://www.deepskyai.com/newsletter/ai-finds-15000-hidden-near-misses](https://www.deepskyai.com/newsletter/ai-finds-15000-hidden-near-misses)
- **Also on Substack:** [https://deepsky294.substack.com/p/ai-finds-15000-hidden-near-misses](https://deepsky294.substack.com/p/ai-finds-15000-hidden-near-misses)
- **Subscribe:** [https://deepsky294.substack.com/subscribe](https://deepsky294.substack.com/subscribe)

*Date (UTC): 2026-03-19*

An AI-powered sweep of radar data has exposed thousands of unreported helicopter close calls at major airports, prompting the FAA to scrap visual separation rules overnight. Meanwhile, the MRO sector braces for a maintenance supercycle as digital twins and agentic AI reshape how fleets are sustained.

##### [FAA Tightens Helicopter Safety Rules After AI Uncovers 15,000 Close Calls](https://www.npr.org/2026/03/18/nx-s1-5752185/faa-airport-safety-rules-helicopters)

The FAA used AI tools to analyze raw radar data and identified over 15,000 previously undetected helicopter-airplane close calls at busy airports between October 2021 and December 2024. In response, the agency issued a new safety order requiring ATC to use radar-based separation for helicopters in approach and departure corridors, ending the longstanding "see and avoid" visual separation method.

##### [The Great Maintenance Crunch: Why 2026 Is the Year of the Aging Fleet](https://oat.aero/2026/03/18/aviation-maintenance-trends-2026-digital-twin/)

With roughly 30,000 aircraft being kept in service 5-7 years longer than planned and 30% of North American AMTs over age 60, the MRO sector faces a capacity crisis. AI-driven digital twins and VR-based training programs are being deployed across Part 145 stations, reportedly reducing unscheduled maintenance events by 35-40% and redefining the aircraft engineer role around data analytics.

##### [GE Aerospace and Palantir Expand Agentic AI Partnership for Military Readiness](https://www.geaerospace.com/news/press-releases/ge-aerospace-and-palantir-expand-partnership-transform-military-aircraft-readiness)

GE Aerospace and Palantir announced a multi-year expansion of their AI partnership, deploying agentic AI across military aviation sustainment and production systems. Starting with USAF T-38 trainer jets, the scope now covers broader MRO, supply chain, and engine production — with AI agents handling fulfillment, sourcing, and maintenance workflow orchestration via Palantir's AIP platform.

##### [MHI and Shield AI Complete Autonomous Drone Flight Tests in Japan](https://www.mhi.com/news/26031702.html)

Mitsubishi Heavy Industries and Shield AI completed autonomous flight demonstrations integrating Shield AI's Hivemind autonomy software into MHI's rapid-prototype drone platform. Two fixed-wing UAVs autonomously demonstrated reinforcement-learning behaviors, coordinated maneuvers, and virtual-target pursuit — with the entire AI-to-flight pipeline completed in just eight weeks.

##### [DroneShield Integrates AI-Powered Radar for Layered Airspace Protection](https://dronelife.com/2026/03/18/droneshield-integrates-robin-radar-for-layered-drone-airspace-protection/)

DroneShield partnered with Robin Radar Systems to add 360-degree 3D radar drone detection to its counter-UAS ecosystem. The integration feeds into DroneShield's DroneSentry-C2 platform powered by SensorFusionAI, fusing radar, RF, and other sensor inputs to create a consolidated operational picture for faster, higher-confidence drone threat response at airports and critical infrastructure.

Why this matters: From rewriting safety rules based on AI-discovered risks to deploying autonomous drones in eight weeks flat, this week shows AI is compressing aviation timelines that used to take years into months — and the regulatory, maintenance, and defense sectors are all moving at once.

If you need help navigating the safe adoption of AI in your flight operation, then we have you covered. Contact us at sam@deepskyai.com or visit deepskyai.com

![image](https://substackcdn.com/image/fetch/$s_!ymjc!,w_1456,c_limit,f_auto,q_auto:good,fl_progressive:steep/https%3A%2F%2Fsubstack-post-media.s3.amazonaws.com%2Fpublic%2Fimages%2F601c49d2-cf6b-45a3-9ce7-7f61cbf66ae1_812x814.png)