The Source Vanishes Before the Article Is Written

On November 27, 2026, the technology outlet Futurism published an investigation into Sports Illustrated. After being questioned about the origin of articles published under the names of authors who did not exist, the biographies and profile images of that author group disappeared from the site within hours. No announcement. No correction. The articles stayed where they were, still indexed, still quietly collecting readers.

I learned my trade by cross-checking things like that. In 2026, aged 17 and interning at a London local paper, I spent three weeks with the financial reports of the Leyton Orient youth academy. I counted 37 sponsorship contracts with unusual refund clauses, money routed through a shell company registered in the British Virgin Islands. I took the reconciliation sheet to my editor. He laughed and said girls usually watch football with emotion. I did not argue. I wrote the piece, published it on my personal blog, and it drew more than 12,000 reads in a single night.

A contract exists only on paper; the money vanished long ago.

Football writes faster than it verifies

Look at the money to see the scale. According to FIFA's global transfer report published in early 2026, clubs worldwide spent 9.63 billion US dollars on international transfers in 2026, the highest figure ever recorded. Fees paid to agents alone set a record at 888.1 million dollars. In January 2026, Chelsea signed Enzo Fernández for 106.8 million pounds, a British transfer record. In August of the same year, Moisés Caicedo joined Chelsea for 115 million pounds, breaking the record just set. Earlier, in 2026, Neymar left Barcelona for Paris Saint-Germain for 222 million euros.

Those numbers are verifiable. They sit in registration filings, annual reports, federation documents, club balance sheets. The problem sits in the layer of content written about them.

Each day, a single Premier League club can generate dozens of headlines. That demand is not met by the number of reporters. It is met by aggregation systems, secondary outlets and reposting accounts. In London, where I live and work, the number of professional football journalists has barely grown in a decade. The number of articles that must be published has multiplied. What fills that gap, I think most people can guess.

Anatomy of a story with no source

Last year, a media company showed me their automated news pipeline, on condition that I not name them. I spent two weeks reading their input-validation code and operating logs.

Each article enters the system as a data object with roughly twelve fields: title, source, type, domain, one-sentence summary, author stance, list of information points, entities mentioned, time sensitivity, source quality. Before any analysis step runs, the system must perform an integrity check.

In one run I read, that check returned negative on almost every field. Empty title. Empty source. Undefined type. Empty information points. Not a single named entity. The only surviving field was the domain label: football.

What matters is how the system handled the rest.

The domain label survives because the classifier runs first and runs independently of the extractor. It is a label, not evidence. But inside software architecture, a label that exists looks exactly like valid data. And when one stage fails, the stages after it do not stop automatically. They keep running on an empty floor.

I call it cascade failure. Empty input, empty output, and if the final layer has no null gate, it fills the void with whatever sounds most plausible.

The verification standard I use in this trade has three tiers. The mandatory tier covers the title, at least three information points with stated sources, and a list of named entities. The important tier covers the publishing source with a date and the author's stance. The supporting tier is quantitative data such as minutes played, transfer fees, wage bills. A story missing the mandatory tier does not get written, no matter how complete the rest looks.

In quality control, this case is called a negative control. The correct output when there is no data is nothing. Not a short piece. Not a piece carrying a note that the source is being verified. Nothing.

Football is the environment where that principle gets violated most often.

The news cycle here is short enough that a transfer rumour only needs to hold for seventy-two hours to collect its reads. Then it expires on its own, nobody corrects it, and nobody remembers. Alongside that, provenance erodes with every repost. A line posted by an agent becomes a line on an aggregator. That line becomes a headline on a secondary site. That headline becomes the source for an analysis piece. By the fourth tier, nobody knows what the first tier was.

The paradox is that the most serious cases in English football have all been recorded on paper. In November 2026, Everton were deducted 10 points for breaching profitability and sustainability rules, reduced to 6 on appeal in February 2026. In March 2026, Nottingham Forest were deducted 4 points. In both cases, what decided the club's fate was not the press release but the balance sheet. The statement said one thing. The books said another. And the books always win.

I once mispronounced the name Perišić, but I am never wrong about what I have witnessed.

In 2026, aged 18, I was the youngest reporter in the press room in Moscow. I mispronounced the name Ivan Perišić three times on camera. A local commentator said flatly that I should go back to a keyboard. It stung, but the reflex I had learned from the Leyton Orient case was still there: open the financial records and read. I found an item labelled match compensation fee worth 2.7 million pounds paid to a sponsor that did not appear in the official accounts. I got the player's name wrong. The number I checked was right.

By 2026, the method had changed. While analysing the 300 million pound rescue package for English football during the pandemic, I found six National League clubs excluded from the support list because of administrative registration errors, a combined loss of 1.4 million pounds. I did not write an exposé. I convened an online meeting, and together with representatives of the six clubs drafted a joint petition to the Department for Culture. Within eight weeks, four clubs got their money back. In the lower leagues, people do not need the spotlight; they need a roof when the rain arrives.

In 2026, in Qatar, I reached 2,154 stadium construction workers. Thirty-eight per cent of them were not receiving the full wage written in their contracts. When a Bangladeshi worker named Rahim had three consecutive months of wages withheld, I did not stop at recording his account. I sat with him and reconciled each pay period against his bank statements, turning testimony into a table that could be independently verified. Every number must have a face. And every face must have a receipt.

The other side of the argument

The defenders of automation have a real case, and I do not want to dismiss it.

Scale is a value. Thanks to automated pipelines, thousands of matches in the National League, in the women's game, in divisions abandoned by local newspapers, now have line-ups, timelines and readers for the first time. Fifteen years ago, a sixth-tier English match did not exist on the internet. Now it does. That is progress, and it came from automation.

The problem lies in where the budget line runs. Writing a thousand articles costs almost nothing. Verifying one costs three weeks, a train ticket, a few phone calls, and sometimes a letter from a lawyer. No business model voluntarily chooses the second option.

So the right argument is not against machines. It is about accountability. A rumour has no byline, no editor, no correction mechanism. It has only a player's name, three days of existence, and an engagement figure nobody audits.

What remains after the article is published

Supporters are the people who pay, but they are usually the last to see the books.

They pay for tickets, for television packages, for shirts. They are the reason the entire 9.63 billion dollar flow exists. But when a story goes up with a confident headline and an empty body, they are the final consumers and the least able to trace anything back.

Football does not end at the ninetieth minute; it runs to the last line of the bank statement.

I am not writing this to conclude that automation is bad. I am writing because I saw a system, on one run, return twelve empty fields and exactly one label: football. And I know, from my own experience, that a domain label has never been proof that a story exists.

A news report only truly exists when someone dares to ask: where is the source?