The quote is real and the photograph is real: statements and images circulating with a public figure's name on them are things they actually said and did
By The Conspiratory EditorsAugust 15, 2026
Where the evidence lands: ContradictedThat quotations, photographs and videos circulating with a named public figure attached are authentic records of what that person said, wore, did or appeared in, and that the presence of a recognisable face or a specific quoted sentence is itself sufficient evidence of authenticity.
Contradicted. Between 10 and 15 August 2026, fact-checkers in three countries rated eight separate artifacts attributed to named public figures, and not one was genuine. The methods barely overlap: a photograph altered to change what a politician was wearing, a photograph of an event that never happened, a quotation invented and attributed to a third party about a target, first-person statements written for people who never said them, and an AI-generated video of two real people in a conversation that never took place. What this file rates is the claim that such material is authentic. It is not. Two things are worth carrying beyond the individual debunks. Several of these were labelled by the people who produced them, one rated by Snopes as Originated as Satire and one posted by an account whose own profile calls its output mostly satire, and the label survived exactly as far as the first repost. And the targeting is not evenly distributed: New America's study of 100 incidents of nonconsensual synthetic intimate imagery aimed at public officials across 14 countries found 92 of them fell between 2022 and 2025, and the burden falls overwhelmingly on women in public life, which is abuse rather than misinformation and is treated here as such.
Believed by: Almost everyone, briefly. These artifacts do not require a conspiratorial worldview: they ask only that a reader accept a plausible-sounding sentence from a person they already have opinions about, which is why they cross political lines and reach audiences no theory would
Published after five days in which the same shape kept recurring across unrelated stories and unrelated countries. The file is built on eight rated artifacts rather than one, because the individual debunks are each unremarkable and the run of them is not. Two limits are worth stating at the outset. The prevalence question is unanswerable from this material, since these surfaced only because three fact-checking organisations happened to pick them up in one week. And the sexualised strand is deliberately named rather than described: the harm literature is cited, the images are not characterised beyond establishing that they were fabricated. source →
The full story
Eight in six days
Between 10 and 15 August 2026, three fact-checking organisations in two countries rated eight separate items attributed to named public figures. A quotation from a White House press secretary. A quotation from an actor about a congresswoman. A photograph of a British prime minister in a slogan T-shirt. A photograph of athletes behind a president. An altered image of a congresswoman. A video of a Senate candidate and a children's entertainer. A retirement statement from a basketball player. A sacking that never happened.
None of it was real. Taken one at a time these are unremarkable, the ordinary background noise of a fact-checker's week. Taken together they are the same operation performed eight times, in politics, in sport, in entertainment, on both sides of the Atlantic, by people with no apparent connection to each other.
Five different ways to make one
What is striking on inspection is how little the techniques share. This is not one tool being passed around.
An image was altered.Full Fact found that the prime minister's “Free Palestine” T-shirt had been added to an existing photograph. Small edit, large implication: it puts a political position on a person's body and leaves them to deny it.
An image was generated whole. The photograph of athletes making an obscene gesture behind the president documents an event that did not occur.
A quotation was invented and given to a third party. The actor Sylvester Stallone was made to say something disparaging about a congresswoman. The target is one person, the mouth belongs to another, and the fabrication attacks the first through the borrowed credibility of the second.
First-person statements were written for people.The press secretary was given a line about the president's intelligence. A basketball player was given more than a hundred words about crying alone and considering leaving the sport.
A video was generated.A Senate nominee and a children's entertainer, in conversation, scored at 100 percent for artificial generation by the Hive Moderation detector.
Five methods, one output. Whatever is driving this is not a technology so much as a demand.
The same account, the same target, four months apart
The press secretary case is the one that changes the picture, because Snopes did not stop at the individual quotation.
It traced the August post to an account it describes as having a history of fabricating statements attributed to public figures. Then it found that the same account had posted a similar fabricated quotation about the same official in April, on the same theme of the president lacking intelligence.
That is not a mistake, a misremembering or a bad screenshot. It is a beat, worked repeatedly against one person over months. And it reframes the rest of this file: some proportion of what looks like scattered noise is a small number of accounts doing this deliberately and often, which is a different problem from a world in which anybody might occasionally invent a quote.
The White House answered Snopes plainly. Asked for comment, it said: “This is fake.” The quotation entered Snopes' most-read list nine days later anyway.
Labelled at the source, unlabelled everywhere else
Two of the eight were satire when they were made, and neither was satire by the time it reached anyone.
Snopes rated the press secretary quotation Originated as Satire. The video of the Senate nominee came from an account whose own profile describes it as building the future of AI media and as “Mostly satire, always entertaining”, with that description sitting one click from the video itself.
Nobody had to strip a watermark or crop a credit. The disclosure was available and the audience did not look, which is worth being precise about because it changes where the failure sits. The producers are not concealing anything. The material still functions as disinformation the instant it leaves the account that made it. So labelling is necessary, plainly insufficient, and mostly protects the person who made the thing rather than the person who reads it.
Who this lands on
One strand of this needs naming rather than describing, and it is not distributed evenly across public life.
New America's study of nonconsensual synthetic intimate imagery aimed at public officials examined 100 documented incidents across 14 countries and found that 92 of them occurred between 2022 and 2025, tracking the arrival of the tools that make them cheap. It sets that alongside a wider evidence base on technology-facilitated abuse of women in public life: a global survey of 14,000 girls and young women in which half reported harassment for voicing political opinions, and a fifth of those went quiet afterwards; research in Kenya and Colombia finding that harassment measurably reduced politically active women's willingness to speak online.
The report quotes the legal scholar Danielle Citron's observation that women function as the canaries in the coal mine for the misuse of new digital tools, and a parliamentarian putting it more bluntly: whatever new thing arrives, it gets used against the women first.
This file therefore treats that material as abuse rather than as misinformation, and handles it accordingly. The doctored image of a congresswoman in this cluster is recorded as having been altered from an original in which she was fully covered, and that is all that is said about it, because establishing that something is fabricated requires nothing further and describing it would do the fabricator's distribution for free.
Why the correction never catches up
The instinct is to say the answer is a denial, and the record in this file is a sustained argument against that.
A fabrication is a finished object. It needs no context, it can be understood in two seconds, and it can be shared without comment. A denial is a claim about an absence. It arrives days later, it reaches a fraction of the people who saw the original, and it cannot be repeated without repeating the thing it corrects. Three words from a White House spokesperson did not stop a quotation reaching a most-read list a week and a half later.
And the hardest cases are the quietest. A false report of an eventcan be checked against a record somewhere: an official announcement, a club statement, a wire service. A fabricated first-person statement about someone's inner life has no counterpart anywhere in the world. The invented retirement statement in this cluster runs to more than a hundred words about private distress, and there is no register anywhere of what a person did not feel. Only the target can deny it, which is precisely the position the fabrication puts them in.
The practical habit, then, is narrow. When a quotation arrives attributed to someone, go and find where they said it, and treat the absence of a source as the answer rather than as an inconvenience. If a sentence in quotation marks cannot be traced to a transcript, a recording, an official channel or a reporter who heard it, the most likely explanation is the simplest one: nobody said it.
What's still unexplained
- What the accounts producing this material are actually for is unestablished in most of these cases. Snopes identified one as a serial fabricator of quotations without establishing a motive, and engagement farming, political intent and ordinary attention-seeking all fit the observable behaviour equally well.
- Whether labelling has any protective effect is genuinely unclear and matters for policy. Two of these carried satire labels at the point of production and functioned as disinformation anyway, which suggests the label protects the producer rather than the audience, but no study cited here measures that directly.
- The true scale is unknown and probably unknowable from fact-checker output. These eight surfaced because three organisations happened to pick them up in one week; the genre's normal condition is to circulate and disappear unrated, so nothing in these feeds supports an estimate of prevalence.
- Whether legal remedies are reaching the sexualised strand is beyond the scope of this file and worth tracking separately. New America's report is framed around policy responses to nonconsensual synthetic intimate imagery, and the gap between documented incidents and available redress is the substance of that argument rather than something settled here.
Point by point
The claim: A direct quotation in quotation marks is a strong claim that somebody would not make up.
What the record shows: People make them up constantly, and one account was caught doing it to the same person twice. Snopes traced the August press secretary quotation to an account it describes as having a history of fabricating statements attributed to public figures, and found that the same account had posted a similar invented quotation about the same official in April, on the same subject. That is not an error or a misremembering. It is a beat, worked repeatedly, and the quotation marks are the cheapest part of the whole production. A sentence in quotes carries an implicit claim that a source exists; nothing about the format requires one to.
The claim: A photograph can be checked by looking at it. If it were doctored, you would see the seams.
What the record shows: Two of the eight were photographs and neither announced itself. In one, a garment carrying a political slogan was added to an image of a serving prime minister, which is a small edit with a large implication, since it places a position on a person's body that they then have to deny. In another, an entire scene was generated: athletes making an obscene gesture at an event where nothing of the kind occurred. Full Fact and Lead Stories settled both by finding the originals and the absence of any corroborating record, not by spotting artefacts. Looking harder at the image is the wrong instinct, because the image is the part that was made carefully.
The claim: If it were fake, the person would deny it, and no denial has been reported.
What the record shows: Denials are issued and they do not travel. The White House answered Snopes' query about the fabricated quotation with 'This is fake', which is about as direct as a denial gets, and the quotation nonetheless entered Snopes' most-read list nine days later. This is the structural asymmetry the whole genre runs on. A fabrication is a complete artifact that needs no context; a denial is a claim about an absence, it arrives later, it reaches a fraction of the original audience, and repeating it necessarily repeats the fabrication. Waiting for a denial is not a verification method, it is a way of being persuaded slowly.
The claim: Most of this is satire, and satire is protected and obvious.
What the record shows: Several of these were satire at the point of production and stopped being satire one share later. Snopes rated the press secretary quotation Originated as Satire. The AI video came from an account whose own profile describes it as building the future of AI media and as mostly satire, always entertaining, with the disclosure sitting one click away on the same page as the video. Nobody had to remove a label; the audience simply did not look. That matters for how blame is allocated. The producers are not hiding, and the material still functions as disinformation the moment it leaves the account that made it, which means labelling is necessary and demonstrably not sufficient.
The claim: This is a general problem that affects public figures equally, so singling out any group is political.
What the record shows: It is not equal, and the imbalance has been counted. New America's study of nonconsensual synthetic intimate imagery aimed at public officials examined 100 documented incidents across 14 countries and found 92 of them occurred between 2022 and 2025, an acceleration tracking the availability of the tools. The report situates this in a wider literature on technology-facilitated abuse of women in public life, including a global survey of 14,000 girls and young women in which half reported harassment for voicing political opinions and a fifth of those self-censored afterwards, and research in Kenya and Colombia finding that harassment measurably reduced politically active women's willingness to speak online. The legal scholar Danielle Citron's formulation, quoted in the report, is that women are the canaries in the coal mine for the misuse of new digital tools. This file treats that strand as abuse rather than as misinformation, which is why it names the pattern and does not describe the images.
The claim: A fabricated statement is easily corrected, so the damage is limited.
What the record shows: It depends entirely on what was fabricated, and the hardest cases are the quiet ones. A false report of an event can be checked against a record: a club statement, an official announcement, a wire service. A fabricated first-person statement about somebody's inner life has no counterpart anywhere, because the only person who can contradict it is the person it was written for. The invented retirement statement in this cluster is exactly that shape, running to more than a hundred words and describing private distress the player never expressed. There is no register of what someone did not feel.
Timeline
- 2026-04An X account later identified by Snopes as having a history of fabricating statements attributed to public figures posts an invented quotation attributed to the White House press secretary, on the theme of the president lacking intelligence on Iran. It is an early instance of what will become a repeat pattern against the same target.
- 2026-08-02The same account posts a second fabricated quotation attributed to the same press secretary, this time reading that the conflict in Iran would be over by now if the president had better intelligence. It spreads across X, Instagram and Facebook without any evidence that she said it.
- 2026-08-06Snopes rates the quotation, using its Originated as Satire rating, and reports that the White House responded to its request for comment with three words: 'This is fake.' The check notes the quotation appears in no official White House material, on no official account, and in no credible news report.
- 2026-08-13An X account carrying the name of a WNBA team posts a fabricated first-person statement of well over a hundred words attributed to a named player, describing distress and a decision to consider retirement. The same account had produced a fabricated report a few days earlier that the team's head coach had been sacked.
- 2026-08-13An account describing itself on its own profile as an AI media platform whose output is 'Mostly satire, always entertaining' posts a video presented as the Democratic Senate nominee in Michigan and a children's entertainer discussing Israel with children. The conversation never happened.
- 2026-08-13Lead Stories rates fake an image of a congresswoman that had been altered from an original in which she is fully covered, and a photograph purporting to show medal-winning athletes making an obscene gesture behind the president. Neither is a record of anything.
- 2026-08-13 to 2026-08-14Lead Stories rates two more: a quotation about the same congresswoman falsely attributed to the actor Sylvester Stallone, and the AI video of the Senate nominee, which the Hive Moderation detector scores at 100 percent for artificial generation.
- 2026-08-14Full Fact rates as edited an image of the British prime minister appearing to wear a 'Free Palestine' T-shirt. The garment was added. It is the eighth artifact in the cluster and the second country.
- 2026-08-15The fabricated press secretary quotation, rated nine days earlier, rises into Snopes' most-read list, alongside a fabricated image connected to a separate story about the same administration. The pattern is now visible in what people are searching for rather than only in what fact-checkers are filing.
Contradicted. Between 10 and 15 August 2026, fact-checkers in three countries rated eight separate artifacts attributed to named public figures, and not one was genuine. The methods barely overlap: a photograph altered to change what a politician was wearing, a photograph of an event that never happened, a quotation invented and attributed to a third party about a target, first-person statements written for people who never said them, and an AI-generated video of two real people in a conversation that never took place. What this file rates is the claim that such material is authentic. It is not. Two things are worth carrying beyond the individual debunks. Several of these were labelled by the people who produced them, one rated by Snopes as Originated as Satire and one posted by an account whose own profile calls its output mostly satire, and the label survived exactly as far as the first repost. And the targeting is not evenly distributed: New America's study of 100 incidents of nonconsensual synthetic intimate imagery aimed at public officials across 14 countries found 92 of them fell between 2022 and 2025, and the burden falls overwhelmingly on women in public life, which is abuse rather than misinformation and is treated here as such.
Reviewed by The Conspiratory Editors · Last reviewed August 15, 2026 · How we rate
Common questions
Is Fabricated quotes and images true?
Contradicted. Between 10 and 15 August 2026, fact-checkers in three countries rated eight separate artifacts attributed to named public figures, and not one was genuine. The methods barely overlap: a photograph altered to change what a politician was wearing, a photograph of an event that never happened, a quotation invented and attributed to a third party about a target, first-person statements written for people who never said them, and an AI-generated video of two real people in a conversation that never took place. What this file rates is the claim that such material is authentic. It is not. Two things are worth carrying beyond the individual debunks. Several of these were labelled by the people who produced them, one rated by Snopes as Originated as Satire and one posted by an account whose own profile calls its output mostly satire, and the label survived exactly as far as the first repost. And the targeting is not evenly distributed: New America's study of 100 incidents of nonconsensual synthetic intimate imagery aimed at public officials across 14 countries found 92 of them fell between 2022 and 2025, and the burden falls overwhelmingly on women in public life, which is abuse rather than misinformation and is treated here as such.
What is Fabricated quotes and images?
A screenshot of a quotation. A photograph of a politician in a T-shirt with a slogan on it. A clip of two familiar faces talking to children. Each arrives already attributed, needing nothing from the reader but recognition of the name. In six days in August 2026, fact-checkers rated eight of these across the United States and Britain and…
What does the evidence show?
People make them up constantly, and one account was caught doing it to the same person twice. Snopes traced the August press secretary quotation to an account it describes as having a history of fabricating statements attributed to public figures, and found that the same account had posted a similar invented quotation about the same offi…
Why do people believe it?
Recognition does the work that verification should. A familiar face or name triggers a sense of knowing the person, and the artifact only has to be consistent with what you already think of them rather than with anything they actually did.
What is still unresolved?
What the accounts producing this material are actually for is unestablished in most of these cases. Snopes identified one as a serial fabricator of quotations without establishing a motive, and engagement farming, political intent and ordinary attention-seeking all fit the observable behaviour equally well.
Sources
- 1.Did Karoline Leavitt say Iran war 'would be over' if 'Trump had better intelligence'?, Snopes (2026)
- 2.Edited image of Andy Burnham in 'Free Palestine' shirt being shared on Facebook, Full Fact (2026)
- 3.Fact Check: FAKE Image Shows AOC In Revealing Sports Bra, Original Has High-Neck, Full Coverage, Lead Stories (2026)
- 4.Fact Check: FAKE Photo Shows Athletes With Medals Flipping The Bird To President Trump's Back, Lead Stories (2026)
- 5.Fact Check: Sylvester Stallone Did NOT Say Talking To AOC Was 'Like Listening To A High School Girl', Lead Stories (2026)
- 6.Fact Check: FAKE Video Shows Abdul El-Sayed, Ms. Rachel Talking To Kids About Israeli 'Colonizers', Lead Stories (2026)
- 7.Fact Check: Sophie Cunningham Did NOT Say She Was Considering Retirement, Lead Stories (2026)
- 8.The Targeting of Female Public Officials (from 'A Weapon Against Women in Politics: Reining in Nonconsensual Synthetic Intimate Imagery'), New America (2025)
- 9.American Sunlight Project, American Sunlight Project (2026)
Embed this verdict on your site
Paste this snippet to show our sourced verdict as a small card, with a link back to the full case file. Free to use.
<iframe src="https://theconspiratory.com/embed/fake-quotes-and-doctored-images" title="The Conspiratory verdict" width="520" height="190" style="border:0;max-width:100%" loading="lazy"></iframe>Help us investigate
This is a living case file. If you spot an error or know evidence we missed, tell us, and weigh in on where you land.
Where do you land?
Cast your read on this one.
Comments
Add your take. Comments are read and approved by a human before they appear, so keep it on topic and civil. Please do not accuse named, living people of crimes.