--- title: Lessons From AI Hacking: Every Model, Every Layer Is Risky description: Two Wiz security researchers share findings from two years of hacking AI infrastructure across five layers: model training, inference, application, AI cloud,... canonical: https://daily.dev/posts/lessons-from-ai-hacking-every-model-every-layer-is-risky-fmqmysx94 twitter:card: summary_large_image twitter:site: @dailydotdev og:type: website og:site_name: daily.dev og:title: Lessons From AI Hacking: Every Model, Every Layer Is Risky | daily.dev og:description: Two Wiz security researchers share findings from two years of hacking AI infrastructure across five layers: model training, inference, application, AI cloud,... og:url: https://daily.dev/posts/lessons-from-ai-hacking-every-model-every-layer-is-risky-fmqmysx94 og:image: https://api.daily.dev/og/posts/fmQmYSX94.png og:image:alt: Lessons From AI Hacking: Every Model, Every Layer Is Risky og:image:width: 1200 og:image:height: 630 og:locale: en --- > ## Documentation Index > Fetch the complete documentation index at: https://daily.dev/llms.txt > Use this file to discover all available pages before exploring further. # Lessons From AI Hacking: Every Model, Every Layer Is Risky **[Dark Reading](https://daily.dev/sources/dr)** · 6 min read · 0 upvotes · 0 comments ## Summary Two Wiz security researchers share findings from two years of hacking AI infrastructure across five layers: model training, inference, application, AI cloud, and hardware. They compromised virtually every major AI platform they targeted, finding vulnerabilities in formats like Pickle, production models like DeepSeek, services like Ollama, vibe-coding platforms, AI cloud providers, and NVIDIA's Triton Inference Server. Their key advice: security teams should focus less on prompt injection and more on fundamental infrastructure vulnerabilities, as nearly every vibe-coded app they examined was hackable within minutes. They recommend continuous security review loops rather than one-time assessments. ## Full article daily.dev links to this article rather than hosting it. Read it at the original source: ## Similar posts on daily.dev - [Securing the AI Stack: From Model to Production](https://daily.dev/posts/securing-the-ai-stack-from-model-to-production-auxghvyo4) · InfoQ · 0 upvotes · 0 comments - [Top cyber threats to your AI systems and infrastructure](https://daily.dev/posts/top-cyber-threats-to-your-ai-systems-and-infrastructure-thdfgbjfx) · CSO Online · 1 upvotes · 0 comments - [The VibeSec Reckoning](https://daily.dev/posts/the-vibesec-reckoning-fy8ffhlpo) · Martin Fowler · 0 upvotes · 0 comments - [Ollama, Nvidia Flaws Put AI Infrastructure at Risk](https://daily.dev/posts/ollama-nvidia-flaws-put-ai-infrastructure-at-risk-skmulrcmh) · Dark Reading · 0 upvotes · 0 comments --- Tags: [#security](https://daily.dev/tags/security), [#vulnerability](https://daily.dev/tags/vulnerability), [#ai-security](https://daily.dev/tags/ai-security), [#prompt-injection](https://daily.dev/tags/prompt-injection), [#ai-infrastructure](https://daily.dev/tags/ai-infrastructure) [View this post on daily.dev](https://daily.dev/posts/lessons-from-ai-hacking-every-model-every-layer-is-risky-fmqmysx94) ```json {"@context":"https://schema.org","@graph":[{"@type":"Organization","@id":"https://daily.dev/#organization","name":"daily.dev","url":"https://daily.dev","logo":{"@type":"ImageObject","url":"https://daily.dev/apple-touch-icon.png","width":180,"height":180},"sameAs":["https://twitter.com/dailydotdev","https://github.com/dailydotdev","https://www.linkedin.com/company/daily-dev-ltd"]},{"@type":"WebSite","@id":"https://daily.dev/#website","url":"https://daily.dev","name":"daily.dev","publisher":{"@id":"https://daily.dev/#organization"},"potentialAction":{"@type":"SearchAction","target":{"@type":"EntryPoint","urlTemplate":"https://daily.dev/search?q={search_term_string}"},"query-input":"required name=search_term_string"}}]} {"@context":"https://schema.org","@type":"TechArticle","headline":"Lessons From AI Hacking: Every Model, Every Layer Is Risky","url":"https://daily.dev/posts/lessons-from-ai-hacking-every-model-every-layer-is-risky-fmqmysx94","mainEntityOfPage":{"@type":"WebPage","@id":"https://daily.dev/posts/lessons-from-ai-hacking-every-model-every-layer-is-risky-fmqmysx94"},"datePublished":"2026-02-23T22:51:32.703Z","dateModified":"2026-02-23T22:52:01.602Z","description":"Two Wiz security researchers share findings from two years of hacking AI infrastructure across five layers: model training, inference, application, AI cloud,...","image":"https://media.daily.dev/image/upload/f_auto,q_auto/v1/posts/3e9ca1e64eeb52951519f1418e71b5f4?_a=AQAEuop","thumbnailUrl":"https://media.daily.dev/image/upload/f_auto,q_auto/v1/posts/3e9ca1e64eeb52951519f1418e71b5f4?_a=AQAEuop","isAccessibleForFree":true,"articleSection":"Dark Reading","inLanguage":"en","publisher":{"@type":"Organization","name":"daily.dev","url":"https://daily.dev","logo":{"@type":"ImageObject","url":"https://daily.dev/apple-touch-icon.png","width":180,"height":180}},"author":{"@type":"Organization","name":"Dark Reading","logo":"https://media.daily.dev/image/upload/t_logo,f_auto/v1/logos/dr","url":"https://daily.dev/sources/dr"},"commentCount":0,"discussionUrl":"https://daily.dev/posts/lessons-from-ai-hacking-every-model-every-layer-is-risky-fmqmysx94","interactionStatistic":[{"@type":"InteractionCounter","interactionType":{"@type":"LikeAction"},"userInteractionCount":0},{"@type":"InteractionCounter","interactionType":{"@type":"CommentAction"},"userInteractionCount":0}],"keywords":"security,vulnerability,ai-security,prompt-injection,ai-infrastructure","timeRequired":"PT6M"} {"@context":"https://schema.org","@type":"BreadcrumbList","itemListElement":[{"@type":"ListItem","position":1,"name":"Home","item":"https://daily.dev"},{"@type":"ListItem","position":2,"name":"Dark Reading","item":"https://daily.dev/sources/dr"},{"@type":"ListItem","position":3,"name":"Lessons From AI Hacking: Every Model, Every Layer Is Risky"}]} ```