It's #FAKE photo of dog protecting baby in earthquake in Venezuela; image was created with artificial intelligence

g1 ? What does the fake photo look like? The photo, which went viral on June 26 on networks such as Facebook, Instagram and X, shows a caramel dog under concrete rubble. He appears protecting a baby with his eyes closed and his face, hands and clothes covered in dirt. The scene also has parts of the construction, with wood, bricks and stones, surrounding the scene. Texts in Portuguese superimposed on the image say: "Image of dog protecting child after earthquake in Venezuela moves the world." At the top, on the left, there is an emoji of a yellow face hugging a red heart. ?? Why is it #FAKE? Fato ou Fake submitted the photo to the artificial intelligence (AI) detector of OpenAI, owner of ChatGPT. The tool pointed out that the content is synthetic. See the scan result (and the infographic below): "Generated using OpenAI tools. The signals below provide evidence supporting this result: SynthID detected. Content credentials not detected." Fato ou Fake submitted the image to OpenAI's artificial intelligence detector, which identified synthetic content Reproduction SynthID is a technology that inserts a watermark to identify synthetic content. This type of seal is incorporated into synthetically manufactured photos. Although imperceptible to humans (who cannot verify the indicator just by observing the scene), the technique is detectable by the system. Regarding the second point ("We did not find a reliable C2PA manifesto from OpenAI"), this is a reference to the acronym for "Coalition for Content Provenance and Authenticity". Created by a non-profit organization based in the United States, the project establishes global technical standards for content origination. The companies that are part of the initiative, including OpenAI itself, provide a type of "birth certificate", or signature, which works as a label to track digital files (images, videos, audios and texts). To try to find out when the photo has been circulating, and the context of the original publication, Fato ou Fake also submitted the content to the Google Fact Check Tools platform. The research showed that the first versions circulated on June 26 in Facebook group posts asking for prayers for Venezuela. The older images contained signs that they were generated by AI, but with serial republication, this warning has been lost. Fato ou Fake has already verified other misleading content about the earthquake in Venezuela. A balance released this Monday (6) by the country's authorities showed that the number of deaths from the double earthquake exceeded 3,500. The tragedy also left 16,740 injured and 17,854 homeless. It's #FAKE image that shows dog protecting baby in earthquake in Venezuela; scene was created with AI Reproduction See also It's #FAKE video of Haaland getting scared when she sees herself with her mouth full in a restaurant mirror' It's #FAKE video of Haaland getting scared when she sees herself with her mouth full in a restaurant mirror VIDEOS: Most viewed

VIDEOS: Fato ou Fake explains SEE other checks carried out by the FATO ou FAKE team Add our WhatsApp number +55 (21) 97305-9827 (after adding the number, send a greeting to be subscribed) GloboPop: click to see videos from the Fato ou Fake stage