Skip to main content
v2026.11,986 entries · CC-BY 4.0

Editorial · CASRAI · AI and ML research outputs

FAA Probes Nationwide Ground Stop as Officials Won’t Rule Out AI Tool

The FAA halted US departures on 13 November 2026; officials cannot rule out a new AI tool, but its maker denies any role and no cause is established.

Published 14 Nov 2026· 4 minute read

Ask CASRAI · free to try

Ask about this story

Ask your first 2 questions free below. Subscribers get 150 a day for $29 a month.

An AI assistant specialized in research administration. It cites the sources behind every answer, labels web answers and says when it can't answer.

Answers draw on CASRAI's guides and dictionary plus the federal and funder documents we index: Federal Register, Grants.gov, Regulations.gov and UKRI.

Works on this site and inside Claude, Cursor and the AI tools you already use.

Everything CASRAI publishes — this page, the dictionary, the guides and the news — stays free to read, with no account and no card.

CASRAI is the reference for research administration — bookmark it for the next question.

On Friday 13 November 2026 the FAA halted departures nationwide for several hours after what it called a communication issue at its air traffic control centres. The cause has not been determined, and officials have said they cannot yet rule out a newly rolled-out AI tool; its developer says it was not involved.

What happened on 13 November

According to Al Jazeera’s report, which draws on AP, Bloomberg, CBS News and flight-tracking data, the FAA ordered the ground stop on the afternoon of 13 November and lifted it at 8:15 pm Eastern. The communication issue affected all 22 air route traffic control centres. FlightAware figures cited by CBS News put delays above 10,000 flights and cancellations at no fewer than 4,000; AP estimated about 1.5 million passengers were affected.

Transportation Secretary Sean Duffy blamed ageing infrastructure and instability in radar and communications systems, saying the agency’s “switches, our routers actually got overloaded with data” and that the system was shut down as a result. He said radar and communications were never fully lost and that the problem has been fixed. Homeland Security Secretary Markwayne Mullin said there was no indication of a threat. Bloomberg reported that the FBI does not believe a cyberattack or criminal activity was responsible and that investigators are examining whether a software update could have been the cause.

Where the AI question comes from

The AI angle rests on timing. The FAA has been adopting SMART (Strategic Management of Airspace, Routes and Trajectories), a tool built by Boston-based Air Space Intelligence (ASI) under an $875 million contract. Al Jazeera reports the FAA announced on 10 November that it would extend the system across all 22 centres. Asked about it, Mullin did not rule it out, saying: “You just can’t rule out anything at this point.” ASI’s statement was direct: “The current ground stop is not based on or a result of the SMART system.”

Those two statements are not in conflict: one declines to exclude a possibility, the other denies it. No reporting to date establishes that SMART or any AI component contributed to the outage.

What SMART is designed to do

In an NPR interview republished by OPB on 1 October, ASI chief executive Phillip Buckendorf described SMART as an advisory tool for traffic managers that draws on more than 200 data sources, including weather, congestion, winds and runway closures, to recommend routing and scheduling. He said: “This is not about automating air traffic control” and “The machine never makes a decision.” Controllers still keep aircraft separated. At that point the system was beginning as a pilot at three Washington, D.C.-area airports.

What this means for researchers and research offices

This is an open investigation, so the useful lessons are about method rather than blame.

  • Co-occurrence is not causation. A new AI deployment and a major outage landing in the same week is a hypothesis to test against logs, change records and the timeline, not a finding. Reports so far point investigators at network and software-update questions first.
  • Advisory tools sit inside larger systems. Evaluating whether an AI component is safe requires evidence about its integration: what data it consumes, what it can change, and what fails if it misbehaves. The NIST AI Risk Management Framework and the safety-case approach both frame assurance at the system level.
  • Incident investigation needs a defined method. Structured approaches such as root cause analysis separate triggering events from contributing conditions, which matters when a single outage has several plausible causes.
  • Public incident records enable research. Resources such as the AI Incident Database only help if reports distinguish confirmed AI involvement from suspected involvement.
  • Organisations deploying AI need a response path. See how an internal AI incident response program is structured.

What is still unknown

The root cause of the outage has not been published, the FAA’s own account of the 13 November event has not identified a trigger, and whether the software update investigators are examining relates to SMART or to anything else has not been stated in the reporting available. This page will be superseded as official findings are released.

Sources

Referenced across the research world

University of Cambridge logoColumbia University logoCrossref logoUniversity of Edinburgh logoHarvard University logoUniversity of Oxford logoPrinceton University logoStanford School of Medicine logoUniversity College London logoORCID logoUniversity of Cambridge logoColumbia University logoCrossref logoUniversity of Edinburgh logoHarvard University logoUniversity of Oxford logoPrinceton University logoStanford School of Medicine logoUniversity College London logoORCID logo
  • University of Cambridge logo
  • Columbia University logo
  • Crossref logo
  • University of Edinburgh logo
  • Harvard University logo
  • University of Oxford logo
  • Princeton University logo
  • Stanford School of Medicine logo
  • University College London logo
  • ORCID logo

View CASRAI adoption →