Skip to main content

How rumors spread during the FAA ground stop

PeakMetrics found about one in six posts about the FAA ground stop carried an unverified claim. Explore the rumors and lessons for comms teams.

Last updated

When officials have few answers, the internet starts offering its own. During the nationwide FAA ground stop, the official explanation remained limited to “telecommunications” through the evening. Online, users filled in the gaps with theories about a new FAA AI tool, cyberattacks, foreign involvement and connections to other headlines.

PeakMetrics analyzed the conversation from the first notice through 8 PM ET. About 17% of posts carried at least one of 14 tracked unverified claims, roughly one in six. For most of those claims, corrections struggled to keep pace.

The biggest rumor attracted some of the weakest correction activity. Meanwhile, unrelated news became material for new theories, and foreign state media and aligned accounts found opportunities to frame the disruption around American weakness.

The biggest rumor was barely corrected

The claim that the FAA’s new AI tool, SMART, caused the ground stop accounted for 31.9% of the rumor conversation. Corrections reached just 3% of that claim’s spread, according to PeakMetrics’ analysis.

Posts claimed the FAA had switched on the tool “today” or “yesterday.” The chronology in the analysis told a different story: the FAA had announced its expansion on November 10, and the tool’s maker said the ground stop was “not based on or a result of the SMART system.”

The timing offered an easy explanation for people looking for a cause. A recent technology announcement and a major disruption were close enough together for users to connect them, even without evidence that one caused the other.

That left a substantial gap between the attention given to the claim and the correction challenging it.

Corrections gained ground unevenly

The cyberattack theory was the second-largest tracked claim, accounting for 17.9% of rumor conversation. It also drew a much stronger corrective response.

Corrections reached 47% of the claim’s spread across the analysis window, rising to 58% during the 6 PM hour. The FBI’s statement and replies from Grok helped circulate those corrections.

Even so, corrective activity remained below the claim’s spread overall.

Only two tracked rumors were outnumbered by corrections: claims of “another 9/11” and false reports that planes had crashed due to the stop. Corrections outnumbered those claims by 2.9 to one and 16 to one, respectively. But the claims represented just 0.2% and 0.03% of rumor conversation.

The clearest correction successes involved two of the smallest rumors. The largest claim remained largely unmatched.

Unrelated headlines became competing explanations

Nearly one in five rumor posts connected the ground stop to other news from the same week. The New Orleans tanker collision accounted for 8.4% of rumor conversation, the Microsoft 365 outage for 7.1%, and NVIDIA CEO news for 4.4%.

These theories offered people a way to make several separate events feel connected. With little detail available about the cause of the ground stop, timing became a substitute for evidence.

Not everything in the tracked conversation had the same status. Alien references, accounting for 8.6%, stemmed from a Stephen King joke and were treated as satire.

That distinction matters. An unverified claim can be a mistaken explanation, deliberate speculation, a joke or an early account that later gains support. Treating all of them as confirmed misinformation would obscure what was actually happening.

Foreign coverage moved quickly and took different forms

Foreign state media began covering the disruption within minutes. Chinese state media accounted for 53% of posts from accounts on PeakMetrics’ monitored adversary list, with CGTN and Xinhua relaying FAA notices during the Chicago-only phase. Those outlets later carried the FBI’s statement that the disruption was not a cyberattack.

Some Russian state media coverage also included officials’ statements that there was no indication of an attack. Being part of the monitored account set did not mean every post promoted an unverified claim.

Elsewhere, particularly in Arabic- and Persian-language commentary, the disruption became a story about “American fragility.” Regime-aligned accounts used it to question US competence and power.

Houthi-aligned accounts showed a more specific coordination signal: three accounts posted identical Arabic text with the same hashtag within one minute. Other posts used a “divine justice” framing or suggested a hidden connection between the ground stop and NVIDIA news.

The same incident was serving several purposes at once: a breaking news story, a source of speculation and material for political messaging.

The biggest theory was largely human

Across the full conversation, about 24% of posts came from accounts rated as likely bots. Their presence varied by narrative.

Likely bots were less represented in the SMART theory than in the conversation overall, even though it was the most-shared unverified claim. The largest rumor therefore cannot be explained primarily by automated activity.

Iran blame was the most bot-heavy narrative: about a third of posts blaming Iran came from accounts rated as likely bots.

For teams monitoring a disruption, both patterns matter. Bot activity can amplify particular claims, while ordinary users can give an unsupported explanation substantial momentum on their own.

Humor was more common than fear

Most posts were neutral or informational, accounting for 65.6% of the conversation. Humor and sarcasm made up 15.5%, compared with 6.2% expressing fear or anxiety. Frustration and anger accounted for 11.5%, while reassurance and support represented 1.2%.

Humor appeared more than twice as often as fear. Some jokes connected the ground stop to Microsoft’s outage, while frustration focused largely on limited official communication in the first hour.

The presence of alarming claims did not mean panic defined the wider conversation. People were also sharing updates, making jokes and expressing annoyance. Understanding that balance helps comms teams respond to the conversation people are actually having.

What this means for comms teams

For comms teams, it’s difficult to keep people informed while the cause of a disruption is still being investigated. But when updates leave basic questions unanswered, speculation has more room to grow.

The FAA conversation shows why it helps to track which explanations are gaining traction alongside the operational response. A correction may be available and still barely register against the claim it addresses. Teams need to know whether their answers are reaching the conversations where people are looking for them.

That means being clear about what is known, what is still being checked and when people can expect another update. It also means paying attention to the claims spreading fastest, so the next communication addresses the questions people are already asking.

Methodology: PeakMetrics analyzed a 10% random sample of social posts about the FAA ground stop across X, Reddit, TikTok, Instagram, Facebook, Threads and Bluesky from 3:31 to 8:00 PM ET on November 13, 2026, using AI classification validated by analyst hand checks.

Sign up for our newsletter

Get the latest updates and publishings from the PeakMetrics investigations team.

Loading the newsletter sign-up form…