Itaú Unibanco · AI used against the institution · severity 1 of 5
In June 2025 sponsored social-media ads used video of Itaú Unibanco board co-chair Pedro Moreira Salles with AI-manipulated audio to promote an investment platform promising up to R$10,000 a day and send viewers to a suspicious link. Itaú said on its #ÉFake blog on 4 June 2025 that the video was a deepfake scam "gerado por inteligência artificial" (generated by artificial intelligence) and that Moreira Salles had no connection to it. g1's Fato ou Fake reported that a detection tool gave a 99.9% probability the audio was AI-generated and counted 220 ads with the same video in Meta's ad library on 6 June. No losses have been reported.
Record last reviewed Oct 21, 2026 · Markdown
Date
Institution
Itaú Unibanco
Country
Brazil
AI system
AI-generated deepfake video and manipulated audio of a bank executive (tool not identified)
AI?
Described as AI by the institution or regulator
The institution, a regulator or a court called the system AI.
Status
Confirmed by the institution
Severity
1 of 5
Impact
No losses or victims reported. · At least 220 ads with the same video ran on Facebook and Instagram, per g1.
What happened?
On 4 June 2025 Itaú Unibanco published a #ÉFake alert saying a video circulating on Facebook improperly used the image of Pedro Moreira Salles, co-chair of its board, to lend credibility to a "miraculous" investment. The bank said the video was generated by artificial intelligence, simulated Moreira Salles making false statements to get users to click a suspicious link, and led to a page imitating the Estadão news portal.[1][2]
g1's Fato ou Fake found the ad running on Instagram, sponsored by an account called "Austrália News", with a caption presenting Moreira Salles as recommending a $100 investment. In the audio the fake executive promised earnings of R$10,000 a day and claimed government backing. The fact-checker ran the video through Hive Moderation, which gave a 99.9% probability that the audio was produced with AI, and found the linked site was anonymous and registered abroad.[2]
On 6 June 2025 Fato ou Fake counted 220 ads with the same video and advertiser in Meta's ad library, targeted at Facebook and Instagram. Meta told g1 the link was unavailable and that activity aimed at deceiving or defrauding others is not permitted on its platforms.[2]
When did it happen?
- Jun 4, 2025
Itaú publishes a #ÉFake alert calling the Moreira Salles video an AI-generated deepfake scam.[1]
- Jun 6, 2025
g1 Fato ou Fake counts 220 ads with the same video in Meta's ad library; Meta says the link is unavailable.[2]
- Jun 7, 2025
g1 publishes its fact-check, reporting a 99.9% AI-generation probability for the audio.[2]
What was the impact?
- Customers: No losses or victims reported.[2]
- Other: At least 220 ads with the same video ran on Facebook and Instagram, per g1.[2]
Severity 1 of 5: Contained executive impersonation aimed at the public: the bank disowned the video and no losses, victim count or regulatory action have been reported.
How did Itaú Unibanco respond?
Itaú said Pedro Moreira Salles "não tem qualquer relação com a publicação" (has no connection to the post), told people to treat such content as suspicious, not to click or share it, and to report it to the social network.[1]
What was the outcome, and did regulators act?
Meta told g1 the link was no longer available. No regulatory or law-enforcement action specific to this video has been reported as of 21 October 2026. Itaú continues to publish deepfake scam alerts; in July 2026 it issued a separate general warning about deepfake and voice-clone scams, listed in the register's responses.[1][2]
Was the Pedro Moreira Salles investment video real?
No. Itaú Unibanco said on 4 June 2025 that the video was a deepfake generated by artificial intelligence and that Moreira Salles had no connection to it, and g1's Fato ou Fake reported a 99.9% probability that the audio was AI-generated.
Was any money lost to the Itaú Moreira Salles deepfake ad?
No losses or victims have been reported. g1 counted 220 ads with the same video on Facebook and Instagram on 6 June 2025, but neither Itaú nor the fact-checker disclosed how many people clicked the link or sent money.
What did Itaú do about the deepfake of its co-chair?
It published a #ÉFake warning on its blog, said the executive had nothing to do with the post, and advised people not to click or share it and to report it to the platform. Meta told g1 the linked page was no longer available.
When AI goes wrong at a bank, the next morning's brief says so.
when one of these regulators moves, the next morning's brief says so · six sourced stories · 7 am PT · free
plus every regulation, bank and agent page update, the morning after · leave any morning