Skip to main content
Low reported severity Media Coverage

Viral images of the US president kissing a White House aide: fact-checks found signs of AI generation, and a White House official called the image fake

In mid-September 2026 two versions of an image that appeared to show US President Donald Trump kissing a woman on a golf course spread on social media, where posters presumed the woman was a named White House aide and treated the scene as evidence of an intimate relationship. Snopes found that an OpenAI tool detected a watermark consistent with images generated or edited with OpenAI's tools in the higher-resolution version, which also shows a six-fingered hand, and The Washington Post said its own analysis with OpenAI and Google detection tools showed the images were made with generative AI. A White House official called the image clearly a fake photo. Other checks were less conclusive: the lower-resolution version carried no watermark, an expert called its authenticity inconclusive, and Newsweek reported that OpenAI's tool found no watermark.

AI system
Unidentified image tool
Occurred
Sept 2026
Reported
15 September 2026
Event location
Unknown
What the AI did
Depicted or impersonated the person
Reported harm
Reputational Harm
Whose AI use
Someone else’s AI use
Setting
Public life
Evidence
AI involvement reported · Causal attribution alleged · 4 sources
10 claims: 1 corroborated, 9 reported. 3 open questions
People reported harmed
At least 1 person

AI system as recorded: A generative AI image tool used to make or edit the circulating images, according to analyses reported by Snopes and The Washington Post; Snopes reports an OpenAI tool detected a watermark consistent with OpenAI's tools and an expert reported a SynthID watermark, while Newsweek reports OpenAI's tool found none (tool not established)

What Happened

What circulated. Snopes reports that two versions of one image, a higher-resolution version and a blurrier lower-resolution version, were shared on X, Threads and Instagram in mid-September 2026. They appear to show the president kissing a woman whose back is turned to the camera on a golf course. The earliest copy Snopes found through reverse image search was the thumbnail of a 14 September YouTube video by a political commentator, who told Snopes his team found the image online, did not create or alter it, and could not confirm who the woman was. Newsweek reports that the images circulated alongside comments treating the encounter as genuine, including allegations of an intimate relationship between the two.

The AI analyses. Snopes reports that an OpenAI verification tool detected a watermark in the higher-resolution version consistent with images generated or edited with OpenAI's tools, and that the woman's hand on the golf club has six fingers. Hany Farid of GetReal told Snopes that the higher-resolution version was derived from the lower-resolution one and carries a SynthID watermark. The Washington Post wrote on 19 September that the images were fakes made with generative AI tools, according to its own analysis using detection tools from OpenAI and Google; only the opening paragraphs of that article could be read.

Contrary and unresolved points. The lower-resolution version had no OpenAI watermark and was too blurry for visible anomalies, and Farid called its authenticity inconclusive; Snopes left the claim unrated. Snopes also reports that OpenAI's tool detected its watermark on a real photo in which only a small portion had been edited with AI, so the watermark alone does not show that the whole scene was generated. Newsweek reported that OpenAI's tool found no SynthID watermark or C2PA record linking the image to OpenAI products, while noting this did not rule out another AI system. Gizmodo reported that AI image detectors gave inconclusive results, some suggesting a real photo; it raised the possibility that someone upscaled the low-resolution version and added the extra finger, but judged it more likely that the clearer AI fake circulated first and others blurred it to hide the AI tells. Who first made the image, and with which tool, is not established.

The response. A White House official told Snopes the image was clearly a fake photo and told Newsweek that it came from Democrats who peddle fake images.

Reported harm

Images that analyses by Snopes and The Washington Post found to bear signs of generative AI showed the president apparently kissing a woman presumed to be a White House aide, and circulated with allegations of an intimate relationship; a White House official called the image fake.

Outcome

Unknown

Snopes, Newsweek, Gizmodo and The Washington Post published checks of the images between 15 and 19 September 2026. A White House official told Snopes and Newsweek the image was clearly a fake photo. Snopes left the claim unrated pending further evidence. No platform action, complaint or legal step concerning the images is reported in the sources read.

What remains unknown

  • Who made the image, with which tool, and whether the lower-resolution version was itself generated or a real photograph later edited.
  • Whether the woman shown is the aide; her face is not visible.
  • Whether the aide or the president took any step beyond the White House official's statement.

What the evidence supports

AI involvement: reported. Two examining parties attribute the images to generative AI on stated grounds. Snopes reports that an OpenAI verification tool detected a watermark consistent with images generated or edited with OpenAI's tools in the higher-resolution version, which also shows a six-fingered hand, and that Hany Farid of GetReal found a SynthID watermark in it. The Washington Post says its own analysis with OpenAI and Google detection tools showed the images were made with generative AI. That generated depiction of the president kissing a presumed aide is what circulated with allegations of an intimate relationship and what the White House objected to. The evidence is not uniform: the lower-resolution version carried no watermark and Farid called its authenticity inconclusive, Newsweek reports that OpenAI's tool found no watermark, and the tool and maker are not established.

10 claims: 1 corroborated, 9 reported. What the statuses mean

Corroborated Images purporting to show the president kissing a White House aide on a golf course spread widely on social media.

Causal attribution. none

  • snopes.com(opens in new tab) supports · English
    The rumor stems from two versions of the image shared on social media in mid-September 2026.
  • washingtonpost.com(opens in new tab) supports · English
    Images purporting to show President Donald Trump kissing White House aide ... on a golf course ricocheted about social media this week, amassing millions of views across online platforms.
  • newsweek.com(opens in new tab) supports · English
    Images claiming to show President Donald Trump kissing ... on a golf course have circulated widely on social media
Reported Snopes reports that an OpenAI tool detected a watermark in the higher-resolution version consistent with images generated or edited with OpenAI's tools, that the woman's hand shows six fingers, and that Hany Farid said this version was derived from the lower-resolution one and carries a SynthID watermark.

Causal attribution. reported

  • snopes.com(opens in new tab) supports · English
    detected a watermark in the higher-resolution version image consistent with those generated or edited using the company's AI tools. Moreover, the image depicts ... holding her golf club with six fingers. ... the high-resolution version was clearly derived from the lower-resolution version and has a SynthID watermark
Reported The Washington Post wrote that the images were fakes made with generative AI tools, according to its own analysis using AI detection tools from OpenAI and Google.

Causal attribution. reported

  • washingtonpost.com(opens in new tab) supports · English
    whipped up using generative artificial intelligence tools, according to a Washington Post analysis using AI detection tools from ChatGPT-maker OpenAI and Google
Reported Newsweek reports that the images circulated alongside comments treating the encounter as genuine, including allegations of an intimate relationship between the president and the aide.

Causal attribution. reported

  • newsweek.com(opens in new tab) supports · English
    The images circulated alongside comments treating the purported encounter as genuine, including allegations that Trump and ... were engaged in an intimate relationship.
Reported A White House official called the image clearly a fake photo and said it came from Democrats who peddle fake images.

Causal attribution. none

  • snopes.com(opens in new tab) supports · English
    a White House official said the image shared online was "clearly a fake photo."
  • newsweek.com(opens in new tab) supports · English
    This is clearly a fake photo. Of course this is coming from Democrats who have nothing of substance so instead they peddle fake images
Reported The earliest copy Snopes found was the thumbnail of a 14 September YouTube video; the commentator who posted it said his team found the image online, did not create or alter it, and could not confirm the woman's identity.

Causal attribution. none

  • snopes.com(opens in new tab) supports · English
    The oldest example that turned up was a thumbnail image for a Sept. 14 YouTube video ... We did not create or alter the underlying image, but we also did not independently verify its original source or that the woman pictured is ...
Reported Snopes reports that the lower-resolution version carried no OpenAI watermark and that Farid called its authenticity inconclusive; Snopes left the claim unrated.

Causal attribution. none

  • snopes.com(opens in new tab) supports · English
    While OpenAI's tool didn't identify a watermark within the lower-resolution image shared across X and Threads ... At this point, I'd say the authenticity is inconclusive. ... we've left this claim unrated
Reported Newsweek reports that OpenAI's verification tool found no SynthID watermark or C2PA record linking the image to OpenAI products, while noting that this did not rule out creation with another AI system.

Causal attribution. none

  • newsweek.com(opens in new tab) supports · English
    found no SynthID watermark or trusted C2PA record linking the image to OpenAI products, but cautioned that this did not rule out its creation with another AI system
Reported Gizmodo reports that AI image detectors gave inconclusive results, some suggesting the image was real; it called upscaling of the low-resolution version possible but judged it more likely that the clearer AI fake circulated first and was later blurred.

Causal attribution. none

  • gizmodo.com(opens in new tab) supports · English
    If you run the image through AI image detectors, you’re going to get a big shrug, with some even suggesting it’s real. ... it’s entirely possible that someone took the low-resolution version and upscaled it to make it clearer, adding the telltale extra finger ... The more likely explanation is that the clearer AI fake was doing the rounds, and people started to add blur to hide the AI tells.
Reported Snopes reports that, in its own test, OpenAI's tool detected the watermark on a real photo in which only a small portion had been edited with AI, even after the edited part was cropped out.

Causal attribution. none

  • snopes.com(opens in new tab) supports · English
    Separately, Snopes tested what the tool would find on a real photo where a small portion had then been edited by AI. OpenAI's tool detected the watermark on such an image, even when the photo was cropped to remove the part that had been edited.

Sources

4 sources inspected. Sources that repeat one account do not corroborate each other.

How the sources were read, and where the events happened

Read in English by the research agent (an AI) on 2026-10-22 (HTTP 200); no translation was involved and no human reviewer read it. Snopes fact-check dated 15 Sep 2026, including its email exchanges with a White House official, the commentator whose thumbnail carried the earliest copy, and Hany Farid. Applies to s1.

Read in English by the research agent (an AI) on 2026-10-22 from an Internet Archive capture of 8 Oct 2026 (the live page timed out); the capture is paywalled and holds only the headline, subtitle and first two paragraphs, which were read. No translation was involved. Applies to s2.

Read in English by the research agent (an AI) on 2026-10-22 (HTTP 200); no translation was involved and no human reviewer read it. Newsweek fact-check dated 15 Sep 2026, including a White House official's statement to Newsweek. Applies to s3.

Read in English by the research agent (an AI) on 2026-10-22 (HTTP 200); no translation was involved. Gizmodo article dated 15 Sep 2026. Applies to s4.

Event countries: Unknown. Affected-person countries: United States. Court countries: Unknown.

The sources do not say where the image was made or first posted; it spread on X, Threads and Instagram, so no event country is recorded. The depicted people are the US president and, as presumed by posters, a White House aide, so their country is the United States. No court proceeding is reported.

Reviewed for publication 2026-10-22: Published as a core case (depicted_or_impersonated): Snopes and The Washington Post each report analyses that found signs of generative AI in images showing the president apparently kissing a presumed White House aide, the images circulated with allegations of an intimate relationship, and a White House official objected that the image was fake. Contrary findings on the lower-resolution version and the watermark are recorded.

People described

The US president and a woman whom social media posters presumed to be a named White House aide, shown in images circulated as evidence of an intimate relationship; the woman's face is not visible and her identity was not confirmed

People reported harmed in this case

At least 1 person

0 AI participants · 1 other person harmed

Additional affected people are described without a reliable count.

One counted person: the president, whose face appears in the image and whose White House objected to it (a White House official called it clearly a fake photo). The woman presumed to be the aide is left uncounted because the sources say her face is not visible and her identity was not confirmed. Views and posts are not counted.

Counted once within this case. The same person may appear in other cases. This count does not establish AI causation.

Cite this case

Compiled per our published methodology: verification statuses, sourcing standards, and corrections process.

APA

NOPE. (2026). Viral images of the US president kissing a White House aide: fact-checks found signs of AI generation, and a White House official called the image fake. AI incidents. https://nope.net/incidents/2026-viral-images-of-us-president-kissing-white-house-aide-fact-checks-find-ai-signs

BibTeX

@misc{2026_viral_images_of_us_president_kissing_white_house_aide_fact_checks_find_ai_signs,
  title = {Viral images of the US president kissing a White House aide: fact-checks found signs of AI generation, and a White House official called the image fake},
  author = {NOPE},
  year = {2026},
  howpublished = {AI incidents},
  url = {https://nope.net/incidents/2026-viral-images-of-us-president-kissing-white-house-aide-fact-checks-find-ai-signs}
}

Related cases

Low Unidentified image tool

Swedish fact-checkers say a clickbait network spread false posts with AI-generated images claiming MP Mohamed Ali confessed to election fraud and was arrested

Between 10 and 15 October 2026 a Swedish-language Facebook page published twelve posts about Mohamed Abdukardir Ali, a Left Party (V) member of the Riksdag, including claims that he had confessed to election fraud and that police had detained him in the middle of a press conference. The fact-checking site Källkritikbyrån reported on 16 October that the claims are false and the images AI-generated, and linked the page through a shared administrator account to BeeUp, a Vietnamese firm that presents itself as a viral marketing agency. SVT Verifierar reported on 17 October that the images claim he confesses to election fraud and has been arrested, and that comment after comment on Facebook reacted with demands such as 'Utvisa honom!' (deport him). The posts built on real reports of suspected election irregularities connected with his personal-vote campaign in Borlänge, which he denies. No source names the image tool, describes how the images were identified as AI-generated, or reports his response to the fabricated posts.

Low Unidentified image tool

California school board race: a flyer carrying an AI disclosure shows a female candidate at a bake sale; she says it manipulated her likeness

In October 2026 the San Dieguito Faculty Association, a California union representing district educators and staff that has endorsed candidate Amy Flicker for a San Dieguito Union High School District trustee seat, accused the campaign of her opponent, Cajon Valley Union School District superintendent David Ray Miyashiro, of using AI 'to generate deceptive, sexist imagery' in a campaign flyer. The flyer, as published by the Daily Mail, shows Flicker behind a table of baked goods labelled 'Non-profit volunteer' beside Miyashiro at a microphone labelled 'District superintendent', and carries the small-print disclosure 'Ad generated or substantially altered using artificial intelligence.' Flicker said that using AI to manipulate her likeness diminished her public service into a condescending stereotype. The campaign's response is not reported, and which parts of the image were generated or altered is not stated.

Low Unidentified image tool

Italian official condemns alleged AI sexualisation of her photograph in a political post

On the eve of the 27-28 September 2026 Chamber by-election in the Reggio Calabria constituency, Roberto Vannacci, leader of Futuro Nazionale, posted on Facebook a campaign image showing Matilde Siracusano, Undersecretary for Relations with Parliament and a Forza Italia deputy, seated at a table beside the centre-right candidate Fabio Roscioli and Senator Claudio Lotito. According to AGI and RaiNews, the white top Siracusano wears under her floral jacket in the original photograph (taken from her Instagram profile, per RaiNews) had disappeared and a much deeper neckline appeared in its place. The Calabrian blog Iacchite', to which Vannacci pointed, said it had published the image on 24 September, that it was not a real photograph but an image produced with artificial intelligence, and that the AI had reconstructed the neck and neckline area; Vannacci reposted it on 25 September without citing the source, according to the blog. Siracusano told Adnkronos that using AI to undress a woman in order to denigrate her was 'squalid'; RaiNews reports her calling it a crime and announcing her intention to file a criminal complaint (querela), while Adnkronos and ANSA say she was weighing one; she told la Repubblica that her lawyers would act. Vannacci replied that his accusers should 'look better' for whom to blame. Ministers and politicians across parties, including Forza Italia leader Antonio Tajani, condemned the post. No complaint filing, investigation or removal of the post is reported in the inspected sources.

Low Unidentified image tool

New Hampshire Senate Republicans ran Meta ads with an AI-altered image of Black state Senate candidate Leonard Bell amid cash and gold chains, outlet reports

Granite Post reported on 5 October 2026 that the New Hampshire Senate Republicans paid Meta to show Facebook and Instagram users an image of Leonard Bell, the Democratic candidate for state Senate District 18 and the only Black candidate for state Senate in New Hampshire this year, laughing in a leather armchair buried in cash and wearing gold chains, a large ring and a watch. According to the outlet, Meta labelled the ad "Digitally created" and its detail page states that the media was created or edited with AI; Bell’s face appears to have been taken from a family snapshot. Bell told the outlet that everything in the image is false other than his face. Citing Meta’s Ad Library, the outlet says three ads from 7 September put the image on New Hampshire screens 39,000 to 50,000 times. The outlet describes the image as a racial stereotype, and a Democratic co-sponsor of New Hampshire’s 2024 deepfake law called it a racist caricature. The Senate Republicans did not respond to questions.

If you or someone you know is struggling, free and confidential support is available. Find a helpline near you at Signpost.