{"id":221027,"date":"2025-08-27T01:15:01","date_gmt":"2025-08-27T06:15:01","guid":{"rendered":"https:\/\/lifeboat.com\/blog\/2025\/08\/new-ai-attack-hides-data-theft-prompts-in-downscaled-images"},"modified":"2025-08-27T01:15:01","modified_gmt":"2025-08-27T06:15:01","slug":"new-ai-attack-hides-data-theft-prompts-in-downscaled-images","status":"publish","type":"post","link":"https:\/\/lifeboat.com\/blog\/2025\/08\/new-ai-attack-hides-data-theft-prompts-in-downscaled-images","title":{"rendered":"New AI attack hides data-theft prompts in downscaled images"},"content":{"rendered":"<p><a class=\"aligncenter blog-photo\" href=\"https:\/\/lifeboat.com\/blog.images\/new-ai-attack-hides-data-theft-prompts-in-downscaled-images.jpg\"><\/a><\/p>\n<p>Researchers have developed a novel attack that steals user data by injecting malicious prompts in images processed by AI systems before delivering them to a large language model.<\/p>\n<p>The method relies on full-resolution images that carry instructions invisible to the human eye but become apparent when the image quality is lowered through resampling algorithms.<\/p>\n<p>Developed by <a href=\"https:\/\/blog.trailofbits.com\/2025\/08\/21\/weaponizing-image-scaling-against-production-ai-systems\/\" target=\"_blank\" rel=\"nofollow noopener\">Trail of Bits researchers<\/a> Kikimora Morozova and Suha Sabi Hussain, the attack builds upon a theory presented in a <a href=\"https:\/\/www.usenix.org\/conference\/usenixsecurity20\/presentation\/quiring\" target=\"_blank\" rel=\"nofollow noopener\">2020 USENIX paper<\/a> by a German university (TU Braunschweig) exploring the possibility of an image-scaling attack in machine learning.<\/p>\n","protected":false},"excerpt":{"rendered":"<p>Researchers have developed a novel attack that steals user data by injecting malicious prompts in images processed by AI systems before delivering them to a large language model. The method relies on full-resolution images that carry instructions invisible to the human eye but become apparent when the image quality is lowered through resampling algorithms. Developed [\u2026]<\/p>\n","protected":false},"author":427,"featured_media":0,"comment_status":"open","ping_status":"closed","sticky":false,"template":"","format":"standard","meta":{"footnotes":""},"categories":[41,6],"tags":[],"class_list":["post-221027","post","type-post","status-publish","format-standard","hentry","category-information-science","category-robotics-ai"],"_links":{"self":[{"href":"https:\/\/lifeboat.com\/blog\/wp-json\/wp\/v2\/posts\/221027","targetHints":{"allow":["GET"]}}],"collection":[{"href":"https:\/\/lifeboat.com\/blog\/wp-json\/wp\/v2\/posts"}],"about":[{"href":"https:\/\/lifeboat.com\/blog\/wp-json\/wp\/v2\/types\/post"}],"author":[{"embeddable":true,"href":"https:\/\/lifeboat.com\/blog\/wp-json\/wp\/v2\/users\/427"}],"replies":[{"embeddable":true,"href":"https:\/\/lifeboat.com\/blog\/wp-json\/wp\/v2\/comments?post=221027"}],"version-history":[{"count":0,"href":"https:\/\/lifeboat.com\/blog\/wp-json\/wp\/v2\/posts\/221027\/revisions"}],"wp:attachment":[{"href":"https:\/\/lifeboat.com\/blog\/wp-json\/wp\/v2\/media?parent=221027"}],"wp:term":[{"taxonomy":"category","embeddable":true,"href":"https:\/\/lifeboat.com\/blog\/wp-json\/wp\/v2\/categories?post=221027"},{"taxonomy":"post_tag","embeddable":true,"href":"https:\/\/lifeboat.com\/blog\/wp-json\/wp\/v2\/tags?post=221027"}],"curies":[{"name":"wp","href":"https:\/\/api.w.org\/{rel}","templated":true}]}}