Anatomy of a fake story: the Simpsons blackout, measured

Series: How minds are moved · Fake news, fact-checks and how to check for yourself · 1 of 7

On Saturday 7 November, at 16:15 UTC, a Brazilian news site asked a question in a headline: did The Simpsons predict a worldwide blackout on 17 November? Six days later, my news engine LSN News had counted 85 headlines from 74 outlets in 6 languages about the same rumour. Only 23 of the 85 called it false in the headline itself. The one fact-checking outlet among my sources wrote about it on 13 November, six days after the first article. My AI agent and I traced the whole path, hour by hour. This is what the anatomy of a fake story looks like when you measure it.

Contents · 11 sections

TL;DR #

• The claim: a Simpsons episode “predicted” a global power and internet blackout on 17 November 2026. No official source supports it. The episode mentions no global blackout and no year 2026.
• The real ingredients: a routine one-minute US Emergency Alert System test scheduled for 17 November, and emergency-kit advice from Brazil’s civil defence the same week.
• First seen in my data in Brazil (Portuguese), then Latin America (Spanish), then Spain, Italy, the Arab world and Greece. 85 headlines, 74 outlets, 23 separate stories, 14 known countries.
• Half the headlines (43 of 85) were questions that never answered themselves. Most of those articles debunk it in the text, but the headline repeats it.
• The story mutated as it travelled: 72 hours without power, a different date, “because of an AI”, a “global cyberattack”, and even a link to the real Microsoft outage.

The claim, and what it was built from #

A good fake rarely starts from nothing. It is assembled from true pieces.

The first true piece is a test. On 16 September the US Federal Communications Commission published a public notice: the Federal Emergency Management Agency (FEMA) will run a nationwide test of the Emergency Alert System on Tuesday 17 November 2026 at 2:20 p.m. Eastern Time. Radio and TV stations will broadcast a message that says, in the notice’s own words, “This is only a test. No action is required by the public.” Trade press describes it as about a minute long and limited to broadcast and cable, with no phone alerts this year. The backup date is 3 December.

The second true piece is advice. In the first week of November, Brazil’s civil defence recommended that households keep water and food for up to three days, 72 hours. According to the Brazilian fact-checker Aos Fatos, the agency called it purely preventive and said there was no sign of imminent danger. Aos Fatos also found two more ingredients being mixed in: a US military resilience programme announced at the end of September, retold online as “the US army is preparing for a 90-day blackout”, and a British preparedness campaign from late October.

The third piece is The Simpsons. The show has a reputation for “predictions”, so any date plus a Simpsons screenshot travels well. Aos Fatos looked at the 2023 episode in question: Lisa’s diary says a storm comes on 17 November, with a drawing of a man and a turkey running from a cloud. In its words, there is no mention of any global blackout or of the year 2026.

Put the three together and you get a story that feels sourced: a government test, an official “prepare a kit” message, and a cartoon that “knew”.

How it travelled: the numbers #

LSN News groups articles about the same event into stories and keeps every headline with its time, language and outlet. My history starts on 7 November, which by luck is the same day the rumour first appears in my data.

day	headlines	headlines calling it false	languages
7 Nov	2	0	Portuguese
8 Nov	3	0	Portuguese, Spanish
9 Nov	11	2	+ English, Italian
10 Nov	20	8	+ Arabic
11 Nov	11	4	Portuguese, Spanish
12 Nov	23	4	+ Greek
13 Nov	15	5	Portuguese, Spanish

The first headline in my data is from dci.com.br, a Brazilian business and lifestyle site: “Os Simpsons previram apagão mundial no dia 17 de novembro? Entenda a teoria que viralizou” (“Did The Simpsons predict a world blackout on 17 November? Understand the theory that went viral”). Terra, one of Brazil’s largest portals, followed five and a half hours later.

By language, it is a Spanish and Portuguese story: 41 Spanish headlines from 38 outlets and 39 Portuguese headlines from 31 outlets, plus 2 in Italian and 1 each in Arabic, Greek and English. By country, where my site registry knows it (55 of the 85 headlines, 14 countries): Brazil 22, Mexico 11, Spain and Colombia 5 each, Ecuador and Argentina 2 each, and one each from the US, Venezuela, Chile, the Dominican Republic, Egypt, Lebanon, Portugal and Saudi Arabia. The other 30 headlines come from sites with no country on file, among them sites in Britain, Greece, Italy and Uruguay.

So the path is clear. It started in Brazil, where the civil-defence advice gave it local weight. It crossed into Spanish through Mexico and South America, where the US alert test is close news. It reached Spain on 11 and 12 November, and Saudi Arabia’s Okaz and a Greek site on the way.

The main story on LSN News has 60 outlets and 57 independent voices. That matters. This was not one article syndicated across a network. Dozens of newsrooms wrote their own version. A rumour that is big in voices, not just in outlets, is one that many editors judged worth a click.

The headline problem #

I read the 85 headlines one by one.

• 43 were a question that did not answer itself: “Vai ter apagão mundial em 17 de novembro?”, “¿Habrá un apagón mundial el 17 de noviembre?”.
• 23 called it false in the headline: “boato” (hoax), “bulo”, “falso”, “desmienten”, “no hay pruebas”, “sin sustento”.
• 19 did neither: “Los Simpson y el apagón del 17 de noviembre: ¿qué pasará?”, or “Entenda” (“understand”).

The articles behind those question headlines mostly explain, further down, that there is no evidence. But a headline is what most people see. On a feed or a search page, “Will there be a world blackout on 17 November?” next to a Simpsons image works as an advertisement for the rumour.

It is most likely not malice but the economics of search: people type “apagão 17 de novembro”, and a headline that matches the query wins the click. The side effect is that the rumour is repeated, with a news brand next to it, dozens of times.

How it mutated #

Every retelling changed something. In my data:

• The date moved. An English-language site in Britain wrote about “The Simpsons’ November 30 Internet Shutdown Prediction”.
• The duration appeared: Spanish outlets wrote about “72 hours without power or internet”, borrowed from the kit advice.
• New causes were added: “¿Habrá un apagón mundial este 17 de noviembre por una IA?” (because of an AI), and a Portuguese-language site asked “Ataque cibernético global?”.
• A real event was pulled in. On 13 November an Ecuadorian paper asked whether the Microsoft Outlook outage “anticipates” the world blackout of 17 November. It answered with what Microsoft said, but the link had been made in the headline.

That last one is a pattern worth remembering. A rumour with a date waiting in the future collects every real incident that happens before the date as “evidence”.

The corrections, and how late they came #

The correction did exist, early. Aos Fatos, a Brazilian fact-checker, published its “no, it is not true” piece on 9 November, two days after the first headline. Aos Fatos is not among the sources LSN News reads, so it is not in my counts.

Inside my data:

• Brazilian mainstream outlets called it a hoax from 9 and 10 November: O Globo (“Entenda a origem do boato”), Veja, a regional civil-defence director quoted as denying it.
• Chile’s BioBioChile called it a “falso rumor” on 10 November.
• Spain’s que.es wrote “El bulo del apagón mundial” on 12 November.
• The only fact-check outlet among my sources that covered it, Maldita.es, published on 13 November at 15:55 UTC, six days after the first article: “no hay pruebas de que se vaya a producir” (there is no evidence it will happen).

And still, on 12 November, the busiest day, only 4 of 23 headlines called it false. The correction was there. It just did not win the headlines.

What it feels like from the inside #

If you saw this on your phone, you probably did not see a lie. You saw a question from a news site you know, a cartoon you like, and a reminder from your own government to keep water in the house. Each piece is reasonable. Together they produce a small, cheap feeling: what if?

That feeling is the product. Sharing “just in case” costs nothing and feels responsible, like warning your family.

What this case can and cannot tell us #

• It shows spread across news sites, not social media. The rumour was surely bigger on WhatsApp, TikTok and Facebook; I do not measure those.
• Headline counts use word patterns. Some headlines are miscounted either way, and a headline can be neutral while the text debunks firmly.
• My history starts on 7 November. The rumour may have circulated online before any news site wrote about it.

How to notice it #

• A date in the future plus a famous “prediction” is a template. The Simpsons, Nostradamus, Baba Vanga: the format is the same.
• Look for the real event inside the rumour. Here it was an alert test and kit advice. Then read what the real source says.
• Treat a question headline as “nobody has confirmed this”. If they had, it would be a statement.
• Watch for mutations: if every version gives a different date, duration or cause, there is no original source.
• Check whether the new “evidence” (an outage, a storm) is linked by anyone except the rumour itself.
• Count voices, and then check if any of them is a primary source. Sixty newsrooms repeating a question is not one government confirming a fact.

Frequently asked questions #

Was there a global blackout planned for 17 November 2026? #

No official source ever said so. What was scheduled for that day was a FEMA test of the US Emergency Alert System at 2:20 p.m. Eastern Time: a test message on radio and TV that says “This is only a test. No action is required by the public.”

Did The Simpsons predict it? #

No. According to Aos Fatos, the 2023 episode shows a storm on 17 November in Lisa’s diary, with no mention of a global blackout or of 2026.

Why did so many news sites write about it? #

Because people were searching for it. Most articles explained there was no evidence, but 43 of 85 headlines were open questions, and only 23 called it false.

How long did the fact-check take? #

A Brazilian fact-checker published two days after the first headline. The only fact-check outlet among my sources published six days after.

The lesson #

This fake was not invented by a villain. It was assembled from a real government test, real safety advice and a real cartoon, and it was carried by news sites that mostly told the truth in the article and repeated the rumour in the headline. That is how most false stories travel today: true pieces, a missing link, and a question mark. The defence is not to distrust everything. It is to find the true pieces, read what their own sources say, and notice that the question in the headline was never answered with a yes.