Independent · Daily · Every regionLatest edition: 21 November 2026
AI WorldwireVerified AI news from every region Subscribe
Latest

Breaking Analysis

FAA Ground Stop Puts New AI Traffic Tool Under Scrutiny

The FAA halted US departures on 13 November. Officials say old telecom gear failed, not the new SMART AI tool, but a probe continues. Plus Australia and the EU.

Nothing in public reporting yet shows that artificial intelligence caused Friday's nationwide US flight halt. But officials were asked, publicly, before any technical findings existed. That is the lesson.

Key takeaways

What happened on 13 November

The FAA halted departures from US airports because of a problem with air traffic control communications; aircraft already airborne could land [2]. FlightAware counted 1,913 cancellations and 10,494 delays by 6:04 p.m. ET, as reported by Visaverge, not independently confirmed [2]. At least 827 flights had been diverted by 7:05 p.m., per Cirium data reported by AP [8][1]. It was the first nationwide ground stop since the January 2023 NOTAM outage, and AP describes it as the longest in 25 years [8][1].

Duffy said the radars and aircraft communications began showing inconsistent data, and that digital equipment continued to work while older analog systems had problems [1]. Officials pointed to overloaded switches and routers in the FAA's aging telecommunications network [1]. AP reports that officials said new technology being added to the system overloaded its antiquated parts; Mullin said the new equipment "was overwhelming the older system", without saying which equipment [8].

The AI question

During the outage, Mullin said the Department of Homeland Security, through its cyber agency CISA, was looking at whether the FAA's SMART system could be connected, while describing the event as a telecommunications issue [3][2]. SMART, short for Strategic Management of Airspace, Routes and Trajectories, uses real-time weather, congestion and flight-path data to help adjust routes, according to the Transportation Department [3][4]. The FAA began rolling it out in September, and specialists are expected to stay involved in overseeing it [3].

Later, Duffy said the problems were not linked to SMART because it is not connected to what failed [8]. AvWeb reports that SMART gives recommendations to controllers and is not connected to the FAA's STARS or ERAM systems, and quotes Duffy: "There's no possible way SMART would have impacted what happened" [1]. The contractor AirSpace Intelligence said the stop was not based on or a result of SMART or its products, and Microsoft said it had no indication its services were related, as reported by Visaverge, not independently confirmed [2]. Mullin said investigators had found no evidence of a cyberattack or AI-related interference, though the investigation was ongoing [1]; AP reports he said the problem appeared to be mechanical [8].

We note the claims side by side because they come from different people at different times: a company denial, an official denial and an open probe. None is a published technical finding. One outlet, Visaverge, also linked a defence contractor to the outage; we could not confirm that and have left it out.

Why the earlier warnings matter

The scrutiny did not start on Friday. On 9 October, lawmakers wrote to Duffy and FAA Administrator Bryan Bedford asking the FAA to pause its use of SMART, which was being rolled out over the Washington-area airports [10]. Later that month, more than 20 members of Congress asked the Government Accountability Office to review SMART and a related system, FMDS [5]. The lawmakers wrote that there had been "little transparency" on how the AI systems underlying SMART were tested [10][5]. The air traffic controllers' union said new technology should complement, not replace, controllers' judgment [5].

Our analysis: the evidence so far points to old infrastructure, not a model error. But the episode shows a trust gap. When a system is deployed in a safety-critical setting with little public testing detail, any outage invites suspicion, and officials end up arguing from assurances. The fix is boring: published test summaries, clear descriptions of what the tool can and cannot touch, and independent audit. Automation bias is the other risk to watch: recommendation-only tools can still steer human decisions.

Elsewhere: Australia and the EU

Australia. On 12 November the government released a draft Best Practice Guide on AI adoption in the workplace, built on three themes (Plan, Enable, Embed) and eight principles that include worker engagement, designing work, AI governance and work health and safety [6]. Consultation runs to 27 January 2027 [6]. The release cites company examples such as Blackmores cutting a data-entry process from 3 days to 10 minutes; these are the government's reported figures, not independent measurements [6].

European Union. The AI Act changes the Parliament approved in June set high-risk obligations to apply from 2 December 2027 for stand-alone systems and 2 August 2028 for those embedded as safety components, and require machine-readable labelling of AI content by 2 December 2026 for systems already on the market [7]. They became law as Regulation (EU) 2026/1744, published in the Official Journal on 24 July 2026 [9]. For safety-critical sectors such as aviation, that timetable matters.

Earlier this week on AI Worldwire

Date Story
10 Nov Claude Fable 5.2 review
11 Nov Australia plans an AI regulator
12 Nov Healthcare AI Pulse
13 Nov AI Weekly Wrap

What to watch next

FAQ

Did AI cause the US ground stop? Officials have not said so. Duffy attributed it to aging telecom equipment and Mullin said no evidence of AI interference had been found, with the probe ongoing [1].

What is the FAA's SMART tool? An AI-assisted traffic management tool that recommends routes and schedules; specialists review its recommendations [3][2].

When do the EU's high-risk AI rules apply? From 2 December 2027 (stand-alone) and 2 August 2028 (embedded in regulated products), under the changes adopted as Regulation (EU) 2026/1744 [7][9].

This article is analysis, not investment or legal advice.

Sources

  1. FAA Rules Out SMART AI System as Cause of Nationwide Ground Stop, AvWeb, 13 November 2026.
  2. FAA SMART AI Vendor and Microsoft Deny Responsibility for Ground Stop, Visaverge, 13 November 2026.
  3. FAA investigating whether new AI technology contributed to nationwide ground stop, KPRC Click2Houston, 13 November 2026.
  4. SMART Fact Sheet, US Department of Transportation, September 2026. Primary.
  5. FAA AI use sparks new reviews into AI safety, workforce shortages, Federal News Network, 16 October 2026.
  6. Draft best practice guide for AI adoption in the workplace released, Minister for Employment and Workplace Relations, Australian Government, 12 November 2026. Primary.
  7. AI Act: EP approves simplification measures and "nudifier" app ban, European Parliament, 16 June 2026. Primary.
  8. Trump administration blames 'overloaded' systems for ground stop after FAA lets flights resume, Associated Press, 13 November 2026.
  9. Regulation (EU) 2026/1744 amending Regulation (EU) 2024/1689, Official Journal of the European Union, 2026. Primary.
  10. Letter to Secretary Duffy and Administrator Bedford on the SMART tool, Members of the US Congress, 9 October 2026. Primary.

AI Worldwire is independent and has no commercial relationship with the companies covered, unless stated. Research and drafting are assisted by AI; every article is reviewed and approved by a human editor before publication. Subscribe to the AI Worldwire newsletter.

Disclosure: the AI system that assisted this article, Claude, is made by Anthropic, one of the companies covered here.

Subscribe to the AI Worldwire newsletter