Quick Answer: The FAA's new SMART AI tool, which stands for Strategic Management of Airspace Routing Systems, was rolled out on the same day the U.S. faced a nationwide ground stop and a communications outage. CNN aviation analyst Mary Schiavo called the timing "too much of a coincidence," but she also said she doesn't know whether the two are linked. The risk she describes, new software struggling to talk to decades-old systems, is well documented.
The FAA rolled out a new AI-assisted traffic tool, and within hours the national airspace was grinding to a halt. Naturally, people asked whether one caused the other. As Mediaite reported, the story spread quickly after CNN aired Schiavo's remarks. Her argument is more useful than the headline, though. It points at a structural problem that matters far more than the calendar.
What the FAA's SMART tool is supposed to do
According to the press release Schiavo read on air, SMART is an AI tool meant to help manage air traffic across the entire National Airspace System. The $875 million price tag is the figure cited in the source. Its specific job is to manage en route traffic, meaning aircraft cruising between terminal areas at altitude, handled by the FAA's en route centers.
That's a sensible target. En route airspace is where congestion builds up quietly, and where a controller can spend hours juggling thousands of flight plans. An AI system that suggests reroutes or spacing adjustments could reduce delays. The FAA says the tool won't affect safety, and nothing in the source suggests otherwise.
Still, a tool's purpose and its integration are two different questions. The part practitioners tend to skip is what the software has to plug into.
Why new software and old systems keep colliding
Here's the core issue. U.S. air traffic control runs on a patchwork of systems that were built in different decades and were never designed to work together. Schiavo described it as "piecemeal, mishmash," with some components going back to radio, telephone, and even floppy disks.
The en route system is a good example. The FAA's En Route Automation Modernization (ERAM) program was meant to replace the aging HOST computer system that managed high-altitude traffic. ERAM was deployed in phases over many years, and the Government Accountability Office repeatedly flagged schedule slips and cost overruns along the way. The lesson isn't that the FAA can't build software. It's that replacing one layer of a live system while the layers around it keep running is hard.
Here's the counter-intuitive part. The common assumption is that a newer system is safer than an older one. In practice, the riskiest moment is often the handoff between old and new, not either system on its own. Each interface is a place where data formats, timing assumptions, or failover logic can disagree. The new system may work perfectly in testing and still fail when it meets an edge case that only exists in the legacy one.
That's why Schiavo's point about interoperability deserves more attention than the coincidence question. It's the mechanism that actually shows up in outage reports.
What a nationwide ground stop actually means
A ground stop is an FAA traffic management action that holds flights on the ground before departure. It can apply to a single airport, a region, or, rarely, the whole country. The FAA uses it when the system can't safely absorb the traffic it's been given, typically because a facility lost capacity, a runway is closed, or a key communications link is down.
Nationwide stops are unusual. They point to a failure in shared infrastructure rather than a local weather event or a single airport problem. When controllers lose reliable communication with aircraft, they can't safely hand flights off between sectors. The response is to stop the flow at the source.
Here's where it gets interesting: a ground stop doesn't tell you the cause. It only tells you the system decided it couldn't operate safely at full volume. The cause could be a hardware fault, a software bug, a bad data feed, a staffing issue, or an unrelated outside event. Anyone who tells you the cause from the stop alone is guessing.
Coincidence or causation? How to read the evidence
Schiavo was careful. She said, "I don't know if they're linked," and she noted that officials hadn't suggested the stoppage involved threats or criminal activity. Her actual position was closer to "the timing is suspicious and the pattern is familiar" than "the AI tool broke the system."
That's the right level of confidence for now. To move past it, you'd want to know:
- Whether the SMART tool was live in operational airspace or only running in a test or advisory mode
- Whether the communications outage involved systems the tool connects to directly
- What the FAA's post-incident report says about the root cause
- Whether similar outages happened on days when no new software was deployed
Until those answers are public, the honest reading is that the two events are correlated in time and the causal link is unestablished. That doesn't make the concern baseless. It means the claim needs evidence.
That said, there's a real catch. Modernization programs often do cause outages, and the FAA's own record makes skeptics understandably wary. The risk is that a legitimate, mundane failure gets labeled as a cover-up, or that a real integration problem gets dismissed as coincidence. Both mistakes are costly.
Precedent: when modernization caused the outage
The strongest argument for Schiavo's concern is recent history. The table below compares two well-documented incidents.
| Incident | Reported cause | Scope and impact |
|---|---|---|
| NOTAM system outage, January 2023 | FAA said a contractor accidentally deleted a file while working on the Notice to Air Missions (NOTAM) database. The agency stated the outage was not a cyberattack. | Nationwide ground stop; more than 11,000 flights delayed, according to widely reported figures |
| Newark-area radar and communications failures, 2025 | Equipment failures at FAA facilities serving Newark Liberty International Airport | Ground stops and sustained delays at Newark, with knock-on effects across the Northeast |
The NOTAM case is the closest parallel. A system maintenance task, not an attack, took down a nationwide notification system. Nobody had to intend harm for the result to be a national shutdown. It's the kind of failure mode you'd expect when old and new components share a critical data path.
Schiavo's broader point is that this pattern has repeated for decades. Each new system is supposed to reduce risk, and each transition briefly increases it. The lesson isn't to stop modernizing. It's to treat every cutover as a risk event with its own rollback plan.
What this means if you fly or work in aviation
If you're a traveler, the practical takeaway is simple: a ground stop is almost always a safety decision, not a sign of imminent danger. The system is protecting itself by slowing down. Check your airline's app, expect delays, and understand that rebooking tools tend to lag behind the actual event.
If you work in aviation, aviation IT, or regulated infrastructure, the lessons are more specific. Test new systems against the legacy interfaces they will actually touch, not just against clean reference data. Stage cutovers so the old system can take back control within minutes. And publish post-incident analysis quickly, because silence fuels exactly the speculation Schiavo was reacting to.
Most people stop here, but the real question is whether the FAA will disclose enough for the public to judge. That's the part to watch.
Frequently Asked Questions
What is the FAA SMART AI tool?
The FAA SMART AI tool is an artificial intelligence system meant to help manage en route air traffic across the National Airspace System. SMART stands for Strategic Management of Airspace Routing Systems. The source reports a cost of $875 million and says the FAA's press release claims it won't affect safety. Its main goal is managing high-altitude traffic between terminal areas.
How do I know if the FAA SMART AI tool caused the ground stop?
You can't know from the timing alone. CNN analyst Mary Schiavo said she doesn't know whether the two events are linked. Confirmation would require the FAA's post-incident report, which should state the root cause, whether SMART was operational, and whether the communications outage involved systems it connects to.
Why do new air traffic control systems cause outages?
New systems often struggle to communicate with older components that were built decades apart. The biggest risk tends to be at the handoff between old and new systems, where data formats or timing assumptions can conflict. Programs like ERAM show how long these transitions can take and how often they slip.
How long can a ground stop last?
Ground stops vary widely. Local stops can last minutes to hours, depending on weather or equipment problems. Nationwide stops are rare and usually last hours, but durations depend on how quickly the underlying fault is found and fixed. Check the FAA's NAS status page for live information.
Is a ground stop a sign of a safety problem?
A ground stop is usually a protective step. It means the system can't safely handle the traffic load at that moment, often due to communications or equipment issues. It doesn't by itself indicate an imminent crash risk, but the underlying cause still deserves a full, public explanation.
Pass this to anyone who's tracking the FAA SMART AI tool story and assuming the timing proves the case. Ask one question first: has the FAA published a root-cause report that names the failed component? Until it does, the cause is open, and your next step is to watch for that report.