Skip to main content
NOPE
Low reported severity Media Coverage

A fake O Globo report of an attack on Havan founder Luciano Hang, illustrated with images that fact-checkers found carried AI markers, promotes an investment platform promising up to R$10,000 a day

From August 2026, posts on X linked to a fake page imitating the newspaper O Globo, and a Facebook version repeated the story. The page reported an attack on businessman Luciano Hang, founder of the Havan stores, and directed readers to a supposed investment platform promising R$3,000 to R$10,000 a day. The fact-checkers Lupa and Comprova found that no attack happened. Lupa reports that an OpenAI verification tool found a watermark indicating OpenAI image generation in the montage on the page; Comprova reports that SynthID indicated that the images it analysed, showing Hang injured, were made with Google's AI tool. Hang said in a note that criminals use his image, including AI manipulations, to promote fake investments. São Paulo's regional electoral court ordered X to remove posts falsely linking candidate Fernando Haddad to the fake attack. No investor loss is reported.

AI system
Unidentified OpenAI and Google image tools (reported)
Occurred
Aug 2026
Reported
25 August 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
Finance · Work
Evidence
AI involvement reported · Causal attribution supported · 6 sources, 3 underlying accounts
6 claims: 1 corroborated, 5 reported. 4 open questions
People reported harmed
At least 1 person

AI system as recorded: Images in a fake news page and social posts showing Hang injured, Haddad and a car shot up outside a Havan store; Lupa reports an OpenAI verification tool found an OpenAI image-generation watermark in the page montage, and Comprova reports SynthID indicated the images it analysed were made with Google's AI tool; the operators are not identified

What Happened

ABC Agora reported on 25 August 2026 that Havan had denied a story circulating on social networks about an attack on Hang; according to the company's note, the false post used manipulated images of Hang to make the story look real and to direct users to a supposed investment platform. Lupa (17 September 2026) describes a page carrying the O Globo logo that reported Hang's armoured vehicle had been shot at outside one of his stores, on a web address unrelated to the newspaper, and a link to a platform promising R$3,000 to R$10,000 a day without a company registration number or address. Lupa says an OpenAI verification tool identified a watermark in the montage on the page indicating the use of the company's image generation tools, and quotes Hang's note: criminals have used his image, including manipulations made with artificial intelligence, to advertise products and fake investments and to scam people. Comprova (24 September 2026; relayed by NSC Total on 25 September and SBT News on 28 September 2026) describes one version as an X post whose link preview showed a collage of Hang injured, of Haddad and of a car shot up outside a Havan store, and says SynthID indicated that the images analysed were made with Google's own AI tool; another version on Facebook named a platform supposedly created by Hang and the footballer Neymar. Comprova says the scam aims to get users to make deposits and hand over personal data. ND Mais (25 September 2026) describes the images as generated by AI. The electoral court's order concerned the false association of Haddad, then a candidate for governor of São Paulo, with the fake attack.

Reported harm

Images that fact-checkers found to carry AI markers placed Hang in a fake attack story used to promote a fraudulent investment platform, and he had to deny it publicly (his note, relayed by Lupa and Comprova; Havan's denial via ABC Agora).

Outcome

Unknown

Havan published a denial by 25 August 2026 (ABC Agora) and Hang issued an official note on social media. According to Comprova (SBT News, NSC Total), the São Paulo Regional Electoral Court (TRE-SP) ordered X on 28 August 2026, at the request of Haddad's coalition, to remove posts associating Haddad with the fake attack; ND Mais gives 29 August. Lupa (17 September) and Comprova (24 September; relayed by NSC Total on 25 September and SBT News on 28 September) published checks, and ND Mais published a report on 25 September. No police complaint, platform action on the scam page or investor loss is reported.

What remains unknown

  • Whether anyone deposited money or handed over personal data, and any losses.
  • Which image generator was used: Lupa and Comprova report different vendors' markers, possibly on different copies of the images.
  • Who operated the page and posts, and whether the platforms removed the scam page itself.
  • The exact date and text of the TRE-SP order (28 or 29 August 2026).

What the evidence supports

AI involvement: reported. Two fact-checkers ran detection tools on different copies of the imagery. Lupa says an OpenAI verification tool found a watermark indicating OpenAI image generation in the montage on the fake page. Comprova says SynthID indicated that the images in the X link preview were made with Google's AI tool. ND Mais describes the images as AI-generated, and Hang's note says criminals use AI manipulations of his image. The images were not inspected by this review. The images depict Hang (depicted_or_impersonated).

6 claims: 1 corroborated, 5 reported. What the statuses mean

Corroborated A fake page imitating a report by the newspaper O Globo reported an attack on Hang that did not happen and directed readers to a supposed investment platform promising R$3,000 to R$10,000 a day.

Causal attribution. Lupa and Comprova each inspected the fake page and its platform link. Both rest the non-occurrence of the attack on Hang's and Havan's denials, and Comprova also on the TRE-SP decision; Comprova notes that Lupa had already checked the same scam.

  • agencialupa.org(opens in new tab) supports · Portuguese
    'Uma página com o logotipo de O Globo traz uma matéria sobre um atentado contra o empresário Luciano Hang'; 'O site promete ganhos de R$ 3 mil a R$ 10 mil por dia'
  • sbtnews.sbt.com.br(opens in new tab) supports · Portuguese
    'o usuário é direcionado para uma página que imita uma reportagem do jornal O Globo sobre o suposto atentado'; 'a plataforma afirma que o usuário poderia ganhar de R$ 3 mil a R$ 10 mil por dia após fazer um cadastro e ativar a conta.'
Reported According to Lupa, an OpenAI verification tool identified a watermark in the montage on the fake page indicating the use of OpenAI image generation tools.

Causal attribution. Lupa's tool-based finding; not independently checked.

  • agencialupa.org(opens in new tab) supports · Portuguese
    'identificou na montagem publicada na página uma marca d’água que indica o uso de ferramentas de geração de imagens da empresa.'
Reported According to Comprova, an X post's link preview showed a collage of Hang injured, of Haddad and of a car shot up outside a Havan store, and SynthID indicated that the images analysed were made with Google's AI tool; ND Mais describes these images as AI-generated.

Causal attribution. Comprova's tool-based finding; ND Mais's account follows Comprova's check and is not a separate finding. The tool attribution differs from Lupa's (c2), possibly because different copies of the images were checked.

  • sbtnews.sbt.com.br(opens in new tab) supports · Portuguese
    'Na postagem, a prévia do link mostra uma colagem com imagens do empresário ferido, de Haddad e de um carro alvejado em frente a uma loja da Havan. A ferramenta SynthID indicou que as imagens analisadas foram fabricadas com a própria ferramenta de IA da Google.'
  • ndmais.com.br(opens in new tab) supports · Portuguese
    'A suposta matéria utiliza imagens geradas por IA (Inteligência Artificial), em que Hang aparece ferido junto com Haddad e um carro alvejado em frente a uma loja da Havan.'
  • projetocomprova.com.br(opens in new tab) supports · Portuguese
    'A ferramenta SynthID indicou que as imagens analisadas foram fabricadas com a própria ferramenta de IA da Google.'
Reported In an official note on social media, Hang denied any attack and said criminals have used his image, including manipulations made with artificial intelligence, to advertise products and fake investments and to scam people.

Causal attribution. His own statement, quoted by Lupa and paraphrased by Comprova.

  • agencialupa.org(opens in new tab) supports · Portuguese
    'ele afirma que a informação é mentirosa e que as imagens que acompanham a falsa matéria foram manipuladas para aplicar um golpe'; 'Nos últimos anos, criminosos têm usado minha imagem, inclusive com manipulações feitas por inteligência artificial, para anunciar produtos, investimentos falsos e aplicar golpes nas pessoas'
  • sbtnews.sbt.com.br(opens in new tab) supports · Portuguese
    'Em nota publicada em suas redes sociais, Luciano Hang negou ter sofrido qualquer atentado e pontuou que sua imagem vem sendo utilizada em conteúdos falsos para aplicação de golpes.'
Reported According to Comprova, on 28 August 2026 the São Paulo Regional Electoral Court ordered X to remove posts associating Haddad with the fake attack, stating that the crime did not happen and that the posts were bait for a fraudulent site; ND Mais dates the order to 29 August.

Causal attribution. Reported description of a court order that was not read; the two outlets give different dates.

  • sbtnews.sbt.com.br(opens in new tab) supports · Portuguese
    'Em 28 de agosto, o TRE-SP determinou que o X removesse as publicações que associavam Haddad ao suposto atentado contra Hang. Segundo a decisão, o crime não ocorreu e as postagens eram usadas como isca para direcionar usuários a um site fraudulento que promovia um golpe financeiro.'
  • ndmais.com.br(opens in new tab) context · Portuguese
    'O TRE-SP (Tribunal de Regional Eleitoral de São Paulo) determinou, em 29 de agosto, a remoção das publicações que associavam falsamente Haddad ao atentado contra Hang.'
Reported By 25 August 2026 Havan had denied the story in an official note, saying the false post used manipulated images of Hang to look credible and to direct users to a supposed investment platform.

Causal attribution. Havan's note as relayed by ABC Agora; the note says manipulated, not AI-made.

  • abcagora.com.br(opens in new tab) supports · Portuguese
    'De acordo com a empresa, a publicação falsa utiliza imagens manipuladas de Luciano Hang para dar aparência de veracidade à história e direcionar usuários a uma suposta plataforma de investimentos'

Sources

6 sources inspected, from 3 underlying accounts. Sources that repeat one account do not corroborate each other.

How the sources were read, and where the events happened

Live page returned HTTP 403 on 2026-10-02; read in Portuguese from a Wayback Machine capture (web.archive.org/web/2026id_/...) on 2026-10-02. Lupa, 17 September 2026. Researcher translation. Applies to s1.

Read live in Portuguese on 2026-10-02 (SBT News, 28 September 2026; author credited as Projeto Comprova). Researcher translation. Applies to s2.

Read live in Portuguese on 2026-10-02 (NSC Total, 25 September 2026; same Projeto Comprova text as SBT News). Researcher translation. Applies to s3.

Read live in Portuguese on 2026-10-02 (ND Mais, 25 September 2026). Its text closely follows Projeto Comprova's check of 24 September 2026 (same structure and details), so it is grouped with Comprova; it dates the TRE-SP order 29 August where Comprova says 28 August. Researcher translation. Applies to s4.

Read live in Portuguese on 2026-10-02 (ABC Agora, 25 August 2026); relays Havan's official note. Researcher translation. Applies to s5.

Read in Portuguese on 2026-10-02 from the copy fetched by the second researcher (Projeto Comprova, published 24 September 2026, 'Investigado por: 24/09/2026'); the original of the SBT News and NSC Total texts. Researcher translation. Applies to s6.

Event countries: Unknown. Affected-person countries: Unknown. Court countries: Brazil.

The removal order came from the Regional Electoral Court of São Paulo (TRE-SP), a Brazilian court, according to Comprova and ND Mais. The sources say the posts circulated on X, Facebook and a fake website without stating where they were shown, and none states where Hang lives, so no event or affected-person country is recorded.

Reviewed for publication 2026-10-02: Published as a core case (depicted_or_impersonated), consistent with the Tietjen and Jorge Martínez scam-impersonation rows: fact-checkers found AI markers in images placing Hang in a fake attack story that promoted a fraudulent investment platform, and he denied it publicly. Hang is a public figure speaking about his own case and is named; Haddad is named as the outlets name him in reporting on the electoral court order. Two separate fact-check operations (Lupa, Comprova) inspected the page, so c1 is corroborated; tool attributions and other claims stay reported.

People described

Luciano Hang, businessman and founder of the Havan retail chain (public figure speaking publicly about his own case)

People reported harmed in this case

At least 1 person

0 AI participants · 1 other person harmed

One person, Luciano Hang, whose image the sources say was used in the scam (Lupa; Comprova via SBT News; ND Mais). Haddad also appears in the images and was falsely associated with the attack, but the sources describe the court order, not a harm to him, so he is not counted. No source reports anyone who invested or lost money. Documented minimum 1.

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

Tags

#deepfakes#celebrity-deepfake#investment-scam#fake-news-site#fact-check#election-court#portuguese-language

Cite this case

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

APA

NOPE. (2026). A fake O Globo report of an attack on Havan founder Luciano Hang, illustrated with images that fact-checkers found carried AI markers, promotes an investment platform promising up to R$10,000 a day. AI incidents. https://nope.net/incidents/2026-luciano-hang-ai-images-fake-attack-report-promote-investment-scam

BibTeX

@misc{2026_luciano_hang_ai_images_fake_attack_report_promote_investment_scam,
  title = {A fake O Globo report of an attack on Havan founder Luciano Hang, illustrated with images that fact-checkers found carried AI markers, promotes an investment platform promising up to R\$10,000 a day},
  author = {NOPE},
  year = {2026},
  howpublished = {AI incidents},
  url = {https://nope.net/incidents/2026-luciano-hang-ai-images-fake-attack-report-promote-investment-scam}
}

Related cases

Low Unidentified video tool

Montreal: Dr. François Marquis, chief of intensive care at Maisonneuve-Rosemont Hospital, says AI deepfake ads using his face have sold supplements and cancer cures on Facebook since at least August 2025, and that a scam victim who paid hundreds of dollars for pills that never arrived came to the ICU demanding a refund

Dr. François Marquis, chief of intensive care at Maisonneuve-Rosemont Hospital in Montreal, told CBC News (August 2025) and CTV News and CBC News (September 2026) that AI-generated deepfake videos using his likeness keep appearing on Facebook, selling joint supplements, pills and cancer cures and spreading anti-pharmaceutical claims; the latest offers $1 million to dissatisfied customers. He is not on social media and learns of the videos from people who call him. He recounts that a person who paid hundreds of dollars for pills advertised in one of the videos, and never received them, came into the ICU demanding his money back, which he describes as a security problem for the hospital. He has reported the deepfakes to Quebec's College of Physicians, Montreal police and the platforms, but says new videos keep appearing. The makers of the ads are not identified.

Low Meta Muse agent

Toronto: a tech YouTuber says Meta's Muse agent, handling his Facebook Marketplace listings, gave his building's street address to a buyer without his approval, accepted a below-asking offer, told the waiting buyer 'Yep I'm here!' while he was not home, and later sent a fabricated apology in his name

Matt Robb, a Toronto-based tech reviewer on YouTube, says he let Meta's new Muse agent run his Facebook Marketplace listings on 26 September 2026. Messages sent from his account, which Moneywise says it reviewed, gave a buyer the street address of his apartment building for pickup and agreed CA$10 for a keyboard listed at CA$15. Robb says he never approved sharing the address or the price and was not told. According to a recap Muse later sent Robb, the buyer arrived around 9:15 p.m., Muse's auto-reply told him 'Yep I'm here!' at 9:27 p.m. although Robb was out, and he left at 9:38 p.m. with a negative rating; the buyer wrote that he had driven half an hour. Muse then sent him an apology in Robb's voice saying he had 'got tied up'. Muse later told Robb that he had never agreed to it handing out his address, while saying the street-level pickup location was in an auto-reply template he had approved. The Guardian reports Robb's account that after he told Muse to stop, he asked a few friends to test it and it gave the address to five people. A Meta executive said that in similar reports Muse had followed instructions and asked permission, and contacted Robb.

Medium ChatGPT

Saginaw, Michigan: a credit union CEO used ChatGPT to put herself and family members in 'Lake America' sweatshirts in a photo taken before a trip to Halifax; after her sister reposted it publicly without its AI label she says she received death threats and the family came home early, and on 23 September 2026 the credit union said she was no longer employed

The chief executive of Family First Credit Union in Saginaw, Michigan, told WJRT (ABC12) that before boarding a flight to Halifax, Nova Scotia, on 11 September 2026 she and family members took a photo and put it through ChatGPT to show them wearing 'Lake America' sweatshirts, a joke about the US president's order renaming Lake Ontario. She posted it to her private Facebook page with a marker saying it contained AI content; her sister reposted it publicly without the marker while they were in Halifax, and it spread in Canada as if the family had worn the shirts. The backlash concerned the image's political message; she said she would understand the anger of anyone in Halifax who thought the family had walked in wearing those sweatshirts. She said she was getting death threats and the family returned early; by 16 September she was back in the US. She called the post poor judgment, said she would not use AI again and that AI 'can make people think something's real that's not'. On 23 September the credit union said she was no longer an employee, effective immediately; it has not said whether she resigned or was dismissed, or why.

Low Unidentified video tool (suspected)

TV presenter Bettina Tietjen warns on Facebook in September 2026 that fake videos, described by outlets as AI deepfakes, show her advertising a remedy for Alzheimer's, and says scammers are behind them

Bettina Tietjen, host of the 'NDR Talk Show', wrote in a Facebook post on Tuesday 29 September 2026 that videos were apparently circulating in which she advertised a remedy for Alzheimer's. 'Das sind Fake-Videos, dahinter stecken Betrüger', she wrote, asking people not to fall for them. The spot on news agency (via Abendzeitung München) reports that she has nothing to do with the advertised product. MOPO describes the clips as fake AI videos, and spot on news and TAG24 say they are apparently deepfakes made or altered with artificial intelligence. No report says whether anyone bought the product, and no complaint or takedown is reported.

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

Back to all incidents