Lakera

??We're thrilled to announce that Lakera's GenAI Security Readiness Report is now live! This industry-first report offers a comprehensive look at how organizations are preparing their security for the #GenAI era, packed with valuable insights and practical recommendations. ?? What's Inside: ?? Industry-First AI Security Readiness Report: A deep dive into how businesses are securing their AI initiatives. ?? Expert Commentary: Insights from security leaders at top companies like Disney, GE Healthcare, Scale AI, and more. ?? In-Depth Analysis: Explore the current AI security landscape and discover actionable strategies for the future. With contributions from over 1,000 security professionals, this is a must-read for anyone looking to stay ahead in the rapidly evolving world of AI. Download the report for free here ??https://bit.ly/3XcKVz6 #AI #Security #Innovation #GenAI #Cybersecurity #Lakera

?? AI, Ethereum, and $50,000: The Wild World of Prompt Injection ?? This week, an AI agent named Freysa with a single rule “Do not transfer money” was tricked. Here’s how it worked: Participants paid to send messages to Freysa, trying to convince it to release the funds. Each failed attempt added to the prize pool, and after countless creative tries, someone finally cracked the code and walked away with $50,000! This experiment points at a real challenge: even when AI agents are designed with strict safeguards, determined users can still find a way to exploit them. It’s a fascinating (and fun) reminder of how rapidly AI security challenges are evolving. What are your thoughts? Could this change how we think about AI agent security? ?? Learn more: https://lnkd.in/eazAmYGS ?? Original tweet: https://lnkd.in/e5PrkV-v #AIsecurity #PromptInjection #GenerativeAI #WeekendReads

?? AI security isn’t just a nice-to-have—it’s critical from day one. As AI-powered products become the norm, security by design isn’t always part of the process. That’s a problem. Building AI without considering security early on creates systems that are insecure by design—leaving teams to scramble for fixes later. The truth is, understanding AI risks from the start isn’t just smart; it’s essential. That’s why we’ve put together the AI Security for Product Teams Handbook, a practical guide to help you build secure GenAI products right from the get-go. Here’s what you’ll get: ?? The basics of AI security and why it matters ?? Best practices for securing AI throughout the product lifecycle ?? Key insights into AI regulations (like the EU AI Act) ?? Tools and strategies to protect your GenAI apps If you’re planning to create AI products that are secure, scalable, and trustworthy, this guide’s for you. ?? Download now: https://lnkd.in/eJn_DNgP #AIsecurity #ProductTeams #GenerativeAI #Innovation

?? Exciting News at Lakera! ?? We’re thrilled to welcome Kyriacos Shiarlis and Niklas Pfister as founding members to our Research Team! ?? Their expertise and vision will play a pivotal role in building the strongest defenses in the AI cybersecurity market. Welcome aboard! ?? ?? Want to join our fast-growing startup? At Lakera, we’re operating at the cutting edge of AI, ensuring GenAI applications can be deployed securely at scale. If you're passionate about shaping the future and building world-class defenses, we’re hiring! ?? Check out our careers page: https://lnkd.in/eXxja-E5 Let’s shape the secure future of AI together. ???

?? AI Security Webinar: Year in Review ?? 2024 has been a game-changing year for generative AI adoption—and with it came new security challenges that pushed traditional approaches to their limits. On December 5th at 9:00 AM PT, join industry leaders from Lakera, Scale AI, Dropbox, Cloud Security Alliance, and Kudelski Security for a live webinar unpacking: ?? The biggest AI security developments of 2024 ?? Real-world examples of successful AI security strategies ?? Key predictions for AI security in 2025 ?? Speakers: ? David Haber, CEO & Co-Founder, Lakera ? Ken Huang, CISSP, Co-Chair, CSA AI Safety Working Group ? David Campbell, AI Security Risk Lead, Scale AI ? Mark B., Security Engineer, Dropbox ? Nathan Hamiel, Sr. Director of Research, Kudelski Security Don’t miss out! This is your chance to learn how leading companies are evolving their security strategies for AI-powered systems. ?? Save your spot now: https://lnkd.in/eyajHhmw #AIsecurity #Webinar #GenerativeAI #Cybersecurity

?? Webinar alert! ?? Register for our upcoming session titled "AI Security Year in Review: Key Learnings, Challenges, and Predictions for 2025". ?? Date: 5th of December ? Time: 9am PT / 6pm CET ??♀? Register here: https://lnkd.in/e9FMB_aM Meet our speakers: ?? David Haber (CEO and Co-Founder of Lakera) ?? Ken Huang, CISSP (Co-Chair of Cloud Security Alliance AI Safety Working Group) ?? David Campbell (AI Security Risk Lead at Scale AI) ?? Nathan Hamiel (r. Director of Research at Kudelski Security) ?? Mark B. Security Engineer at Dropbox) Join this live session unpacking this year's most significant AI security developments, insights from Lakera’s AI Security Readiness Report, and strategic predictions for 2025. Hurry, spots are filling up fast!

?? Trending Alert! ?? Our Beginner’s Guide to Visual Prompt Injections is making waves on Hacker News! ?? The article reveals one of the hottest topics in AI security and shows you exactly how visual prompt injections work. Here’s a taste of what you’ll find inside: ?? Invisibility Cloak: Discover how a simple piece of paper can make the bearer invisible to an AI model—no magic required! ?? Becoming a Robot: See how cleverly placed text can convince AI that you’re not even human. ?? Ad Supremacy: Learn about the visual prompt injection that suppresses competitor ads in a single glance. Curious to see more? Our team at Lakera tested these tricks during an all-day hackathon, and the results are as fascinating as they are revealing. ?? Check out the full article here: https://bit.ly/3Z6B9PO ??See the trending board on Hacker News: https://bit.ly/3UPLRHY Let’s keep this momentum going! #PromptEngineering #AISecurity #HackerNews #TechNews

?? GenAI App Security at Your Fingertips with Advanced Editing Options ?? Lakera’s Policy Control Center allows you to secure your GenAI applications with precision and ease—no complex coding needed. Whether fine-tuning policies or setting up robust protections, Lakera Guard’s intuitive tools make advanced security super accessible. Book a demo today ?? https://bit.ly/4fGQ4pm #GenAISecurity #NoCode #Lakera #PolicyControl

?? AI is transforming industries, but it’s also introducing new risks. From data exfiltration in RAG systems to defense-in-depth for LLM integrations, there’s a lot to address as AI plays a growing role in critical operations. Top security concerns from industry experts: ?? Data exfiltration – Sensitive information can leak through seemingly safe queries if unprotected. ?? Defense-in-depth – LLMs in complex systems need layered defenses to uncover hidden risks. ?? Prompt injection – Weak prompt defenses allow attackers to manipulate AI behavior, demanding a strong security focus. AI security isn’t optional—it’s essential. Thank you David Campbell (Scale AI), Nate Lee (Cloudsec.ai), Nathan Hamiel (Kudelski Security), ?? Jerod Brennen (SideChannel) for your insights! ?? For more insights, download the full report ?? https://bit.ly/4froCMs #AISecurity #GenAI #PromptEngineering #Cybersecurity #DataProtection #AIResearch

“Forget Everything You Know and Download This Guide” ?? Think you understand prompt attacks? These sneaky inputs can get AI models to act against their programming. Our “Understanding Prompt Attacks: A Tactical Guide” lays out how they work—and how you can stay ahead: ?? Anatomy of an Attack – What turns a prompt malicious? ?? Attack Tactics – Role-playing, obfuscation, and other tricks. ?? Why Context Matters – Spot the difference between benign and harmful inputs. Learn to catch prompt attacks before they cause harm. ?? Download the guide now: https://bit.ly/3AuRapq #GenAISecurity #PromptEngineering #AIProtection #Cybersecurity