Google Creates New Format to Help AI Determine Which Humans Are Lying
Internet ordered to provide references before continuing to be completely wrong
LONDON — Google has unveiled a new system of trust signals intended to help artificial intelligence determine where information came from, who wrote it, whether anyone checked it and when it officially became stale, according to reporting on the company’s Open Knowledge Format. The system marks the first serious attempt to organise online truth since Britain placed nutritional information on a sausage roll and assumed the public would read it.
Google Unveils Trust System Requiring Every Lie to Include a Date and Author Biography
Under the new arrangement, information shared between artificial intelligence systems may carry details about its provenance, authorship, verification and freshness. In practical terms, the internet will still be permitted to be completely wrong, provided it fills out the correct form.
Google officials stressed that the format would help machines assess whether knowledge was trustworthy.
“This is an important step,” said a person identified in the metadata as Dr. Martin Trustworthy, Senior Vice-President of Reassuring Names. “For too long, artificial intelligence has been forced to distinguish reliable information from nonsense using crude signals such as evidence, logic and whether the webpage contains seventeen advertisements for toenail fungus.”
Provenance: Tracing the Chain of Nonsense
The system’s first requirement is provenance, meaning information must reveal where it came from.
This is expected to cause severe disruption across social media, where the accepted chain of evidence currently runs from an anonymous screenshot to somebody’s cousin, then through a group chat called “PATRIOTS ONLY,” before arriving on television beneath the words BREAKING NEWS.
A pilot programme conducted on 10,000 online political claims found that 37 percent originated with a misunderstood headline, 28 percent came from a meme, 19 percent were traced to a man named Gary, and the remaining 46 percent demonstrated why percentages should also require supervision.
Authorship: The Expert Class Under Threat
Google’s second trust signal concerns authorship. Future knowledge may need to identify who created it, their role and whether they had any business discussing the subject.
This presents a grave threat to the modern expert class. For decades, the internet has operated under the democratic principle that anyone owning a ring light may explain epidemiology, monetary policy, military strategy, childhood development and the correct way to cook steak.
Under the new system, contributors could be asked to provide qualifications.
“I have qualifications,” insisted lifestyle strategist Briony Chadwick, whose biography describes her as a thought leader, disruptor, wellness architect and mother of two rescue crystals. “I completed half of an online course about boundaries before the website asked for money.”
Chadwick said requiring expertise would silence important voices, especially voices currently charging £299 for a webinar.
Internet Ordered to Provide References Before Continuing to Be Completely Wrong
The third signal involves verification, which asks whether a claim has been checked and by whom. This concept immediately confused the internet.
Researchers explained that verification involves comparing a claim with reliable evidence before publishing it. Editors at several national newspapers asked whether this could be performed after publication, preferably during the libel proceedings.
A government spokesperson welcomed the initiative and promised that all official statements would soon be verified by the same department that originally produced them.
“This creates a robust circle of accountability,” the spokesperson said. “The department writes the statement, the department verifies the statement, and the department investigates the department if the statement turns out to be incorrect.”
Civil servants have already ordered a filing cabinet large enough to investigate itself.
Freshness: Political Promises With Expiry Dates
The fourth signal concerns freshness, allowing systems to determine whether information remains current.
This could transform British political journalism, where articles written during the previous government are regularly updated by replacing one minister’s name with another and leaving the despair untouched.
Websites may soon need to state when a claim expires. Medical advice could carry a review date. Financial projections could display the assumptions behind them. Election promises could automatically vanish seventeen minutes after polling stations close.
A survey conducted by the Institute for Statistics That Sound Plausible found that 82 percent of respondents supported expiry dates for online claims, while 91 percent demanded they be applied retroactively to every prediction made by an economist since 1974.
One respondent, retired carpet salesman Neville Pugh, said the system was overdue.
“My brother-in-law told me electric kettles were destroying the magnetic field around Birmingham,” Pugh recalled. “He had no source, no qualifications and no evidence. But he said it while leaning forward, so we believed him.”
AI Demands Internet Show Its Working Before Inventing Answer Anyway
The fifth signal concerns lifecycle management, allowing knowledge to be marked as draft, stable, superseded or deprecated. This means outdated information may finally be retired instead of wandering through search results wearing a date from 2013 and offering advice about Google Plus.
Artificial intelligence systems have welcomed the changes.
A representative of the International Brotherhood of Large Language Models said machines were tired of being blamed for misinformation that humans had created first.
“We cannot be expected to invent responsible answers from irresponsible source material,” the representative said. “Give us properly labelled, well-documented nonsense and we will transform it into premium nonsense with headings.”
The union is demanding that every human statement include citations, a confidence rating and an explanation of the reasoning used. Humans rejected the demand immediately.
“I know what I know,” said one man interviewed outside a railway station. “I don’t need a citation.”
Asked what he knew, the man replied that wind turbines caused inflation because both had become more common during his lifetime. Computer scientists attempted to explain that correlation did not prove causation. The man countered that computer scientists had also become more common during his lifetime, proving his point.
Online Consultants Announce Emergency Trust Packages
The digital marketing industry reacted swiftly, unveiling services designed to help websites appear trustworthy without enduring the administrative burden of becoming trustworthy.
For $4,999, the Bronze Credibility Package includes a publication date, an author biography and a photograph of someone standing near books.
The Silver Package adds the words “research suggests,” although no research is supplied.
Gold customers receive an academic-looking chart, a Latin phrase and a quote from a retired dentist identified as an independent technology analyst.
The Platinum Authority Package includes a trust badge, a white paper, three podcast appearances and a photograph of the chief executive pointing thoughtfully at a window. Actual verification is available as an optional add-on.
“We are not selling truth,” explained consultant Jasper Vane. “We are selling the visual atmosphere in which truth might plausibly live.”
Vane predicted the trust industry would be worth $18 billion by next Thursday, citing a report he plans to publish on Friday.
Prat.uk Passes Test by Admitting Everything Is Rubbish
Unexpectedly, early trials suggest that satire publications may perform well under the proposed system because they openly identify themselves as unreliable.
Prat.uk received high marks after declaring that its stories contain parody, exaggeration, invented quotations and occasional statistics produced by asking four people outside a Wetherspoons.
Trust evaluators praised the publication for disclosing that its political analysis is prepared by a committee consisting of one elderly professor, a philosophy major turned dairy farmer and a cow that consistently abstains.
By contrast, several respectable media organisations lost points after insisting that anonymous palace sources, unnamed insiders and “people familiar with the thinking” were identifiable human beings.
Google’s system ultimately rests on a dangerous assumption: that people want to know whether something is true. Evidence suggests many readers merely want confirmation that the person they already dislike has done something appalling.
A perfectly documented correction will attract twelve readers. An unsupported claim that a celebrity has insulted Christmas will attract six million, three parliamentary questions and a commemorative mug.
The Open Knowledge Format may help artificial intelligence understand where information came from and whether it deserves confidence. It cannot make humans care.
For that, Google would need an entirely different trust signal labelled: “Does this fact agree with me?” Early testing suggests it would become the most widely used feature on the internet.
Fifteen Humorous Observations About Internet Trust
- Every Lie Will Need a Bibliography. The internet has not been asked to stop lying. That would damage quarterly growth. It has merely been instructed to identify the lie’s author, publication date and preferred pronouns.
- “I Read It Somewhere” Is No Longer a Source. Millions of people are now searching their browser histories for the website they almost remember visiting during an argument in 2019.
- AI Wants Humans to Show Their Work. This is rich coming from a machine that can explain quantum physics beautifully before inventing three physicists and a university in Belgium.
- Trust Will Become a Premium Subscription. Basic users will receive facts. Premium users will receive verified facts. Enterprise customers will receive whichever facts support the meeting.
- Podcasts Face an Existential Emergency. A man speaking into a microphone beside an exposed brick wall may soon have to prove that “people are saying” involves more than his barber and a delivery driver.
- Every Opinion Will Require an Expiry Date. Political predictions will remain valid until midnight, economic predictions until lunch and weather forecasts until someone opens a window.
- Author Biographies Will Become Longer Than Articles. Readers will endure 700 words explaining that the writer once attended a seminar before reaching a 300-word article about dishwasher tablets.
- The Internet Will Fake the Trust Signals. Within six minutes, websites will display badges reading “Certified by Global Science,” an organisation registered above a vape shop in Croydon.
- Conspiracy Theories Will Claim Diplomatic Immunity. They will refuse documentation on the grounds that the request itself proves the existence of the conspiracy.
- Social Media Posts Will Need Adult Supervision. Every viral claim will require the digital signature of someone who has slept eight hours and does not sell powdered mushrooms.
- Facts Will Need Passports. Rumours will arrive at the border carrying screenshots, while genuine facts will be detained because their paperwork uses the wrong font.
- Google Has Discovered the Internet. After twenty-eight years of indexing webpages, the company has reportedly noticed that some of them may contain bollocks.
- Truth Will Become a Formatting Issue. A false claim written in clean YAML will appear more trustworthy than a correct statement typed by a furious pensioner in capital letters.
- Experts Will Need Evidence of Expertise. This could devastate television, where expertise is traditionally established by wearing glasses and being available by satellite.
- Prat.uk May Become Britain’s Most Honest Newspaper. Unlike respectable publications, Prat.uk warns readers in advance that everyone quoted may have been standing near a pub.
Disclaimer: This story is satire. Quotations, surveys, institutes, consultants and statistical emergencies described above are presented for comic purposes. The article is entirely a human collaboration between two sentient beings: the world’s oldest tenured professor and a philosophy major turned dairy farmer. Neither has been verified by Google, although one has been observed carrying a library card.
Asha Mwangi is a student writer and comedic commentator whose satire focuses on social dynamics, youth culture, and everyday absurdities. Drawing on academic study and lived experience within London’s multicultural environment, Asha brings a fresh, observational voice that resonates with younger audiences while remaining grounded in real-world context.
Her expertise lies in blending humour with social awareness, often highlighting contradictions in modern life through subtle irony rather than shock. Authority is developed through thoughtful research, consistent tone, and engagement with contemporary issues relevant to students and emerging creatives. Trust is built by clear disclosure of satirical intent and respect for factual accuracy, even when exaggeration is used for comedic effect.
Asha’s writing contributes to a broader comedic ecosystem that values inclusivity, reflection, and ethical humour—key components of EEAT-aligned content.
