ActiveFence

𝗛𝗼𝘄 𝗔𝗰𝘁𝗶𝘃𝗲𝗙𝗲𝗻𝗰𝗲 𝗛𝗲𝗹𝗽𝘀 𝗔𝗺𝗮𝘇𝗼𝗻 𝗕𝘂𝗶𝗹𝗱 𝗦𝗮𝗳𝗲𝗿 𝗠𝗼𝗱𝗲𝗹𝘀❓ As AI models become more powerful, ensuring their safety, trustworthiness, and resilience is more critical than ever. Amazon Web Services (AWS)’s Nova models, powering text, image, and video generation, underwent extensive red teaming to identify and mitigate vulnerabilities before deployment. As a key partner in this process, ActiveFence’s AI Safety&Security red team helped strengthen Nova’s defenses, ensuring safer AI that meets the highest industry standards. 𝗔𝗜 𝘀𝗮𝗳𝗲𝘁𝘆 𝗶𝘀𝗻’𝘁 𝗷𝘂𝘀𝘁 𝗮 𝗰𝗵𝗮𝗹𝗹𝗲𝗻𝗴𝗲- 𝗶𝘁’𝘀 𝗮 𝗿𝗲𝘀𝗽𝗼𝗻𝘀𝗶𝗯𝗶𝗹𝗶𝘁𝘆. 𝗟𝗲𝘁’𝘀 𝗯𝘂𝗶𝗹𝗱 𝗔𝗜 𝘄𝗲 𝗰𝗮𝗻 𝘁𝗿𝘂𝘀𝘁. 𝘍𝘰𝘳 𝘢 𝘥𝘦𝘦𝘱𝘦𝘳 𝘥𝘪𝘷𝘦 𝘪𝘯𝘵𝘰 𝘰𝘶𝘳 𝘳𝘦𝘥 𝘵𝘦𝘢𝘮𝘪𝘯𝘨 𝘢𝘱𝘱𝘳𝘰𝘢𝘤𝘩 𝘢𝘯𝘥 𝘪𝘯𝘴𝘪𝘨𝘩𝘵𝘴 𝘧𝘳𝘰𝘮 𝘰𝘶𝘳 𝘤𝘰𝘭𝘭𝘢𝘣𝘰𝘳𝘢𝘵𝘪𝘰𝘯 𝘸𝘪𝘵𝘩 𝘈𝘞𝘚, 𝘳𝘦𝘢𝘥 𝘵𝘩𝘦 𝘧𝘶𝘭𝘭 𝘣𝘭𝘰𝘨 𝘩𝘦𝘳𝘦 👉 https://lnkd.in/dJT6Fdyj

𝗧𝗵𝗲 𝗥𝗲𝗮𝗹 𝗟𝗮𝘀𝘁 𝗠𝗶𝗹𝗲: 𝗗𝗲-𝗥𝗶𝘀𝗸𝗶𝗻𝗴 𝗚𝗲𝗻𝗲𝗿𝗮𝘁𝗶𝘃𝗲 𝗔𝗜 𝗶𝗻 𝗣𝗿𝗼𝗱𝘂𝗰𝘁𝗶𝗼𝗻 At NVIDIA GTC, our Co-Founder & CTO, Iftach Orr, took the stage to address one of the biggest challenges in #AI today: 𝗵𝗼𝘄 𝘁𝗼 𝘀𝗮𝗳𝗲𝗹𝘆 𝗯𝗿𝗶𝗻𝗴 𝗚𝗲𝗻𝗲𝗿𝗮𝘁𝗶𝘃𝗲 𝗔𝗜 𝗶𝗻𝘁𝗼 𝗽𝗿𝗼𝗱𝘂𝗰𝘁𝗶𝗼𝗻 while mitigating real-world risks. 𝗔𝘁 ActiveFence, 𝘄𝗲 𝗰𝗼𝗻𝘀𝗶𝗱𝗲𝗿 𝘁𝗵𝗿𝗲𝗲 𝗰𝗿𝗶𝘁𝗶𝗰𝗮𝗹 𝗮𝘀𝗽𝗲𝗰𝘁𝘀: 🔹 𝗨𝘀𝗲𝗿 𝗿𝗶𝘀𝗸- Even with verification, threat actors will attempt to bypass safeguards. 🔹 𝗠𝗼𝗱𝗲𝗹 𝗿𝗶𝘀𝗸- Hallucinations, harmful content, aggression, and off-topic responses. 🔹 𝗘𝗻𝘃𝗶𝗿𝗼𝗻𝗺𝗲𝗻𝘁 𝗿𝗶𝘀𝗸- AI systems interacting with external tools or proprietary data. 𝗪𝗵𝗮𝘁 𝘄𝗲’𝘃𝗲 𝗹𝗲𝗮𝗿𝗻𝗲𝗱 𝗳𝗿𝗼𝗺 𝘄𝗼𝗿𝗸𝗶𝗻𝗴 𝘄𝗶𝘁𝗵 𝘁𝗼𝗽 𝗳𝗼𝘂𝗻𝗱𝗮𝘁𝗶𝗼𝗻 𝗺𝗼𝗱𝗲𝗹𝘀: ✅ 𝗣𝗲𝗿𝗳𝗲𝗰𝘁 𝘆𝗼𝘂𝗿 𝗽𝗼𝗹𝗶𝗰𝘆- Be specific. It’s easy to say “don’t be hateful,” but defining what that means is where the real work happens. ✅ 𝗜𝗺𝗽𝗹𝗲𝗺𝗲𝗻𝘁 𝗿𝗶𝗴𝗼𝗿𝗼𝘂𝘀 𝗲𝘃𝗮𝗹𝘂𝗮𝘁𝗶𝗼𝗻𝘀- Grey-area test cases refine guidelines and uncover blind spots. ✅ 𝗥𝗲𝗱 𝘁𝗲𝗮𝗺𝗶𝗻𝗴 𝗶𝘀 𝗮 𝗺𝘂𝘀𝘁- Automated tools help, but manual testing is essential. External vendors bring unbiased insights and ensure testing isn’t influenced by internal pressures. ✅ 𝗖𝘂𝗿𝗮𝘁𝗲 𝘁𝗿𝗮𝗶𝗻𝗶𝗻𝗴 𝗱𝗮𝘁𝗮 𝗰𝗮𝗿𝗲𝗳𝘂𝗹𝗹𝘆- Time stamping datasets helps adapt as risks evolve over time. ✅ 𝗠𝗼𝗻𝗶𝘁𝗼𝗿𝗶𝗻𝗴 & 𝗼𝗯𝘀𝗲𝗿𝘃𝗮𝗯𝗶𝗹𝗶𝘁𝘆- AI safety doesn’t stop in the lab, real-world tracking is key to adjusting guardrails as new risks emerge. Want to learn more about making your AI safer? Contact us 👉 https://lnkd.in/dMr5PC-t

“Why Red Teaming is Critical for AI Security - with Tomer Poran of ActiveFence” https://lnkd.in/d_MacVm Red teaming is becoming a critical component of AI security and governance. Tomer Poran, Chief Evangelist and VP of Strategy at ActiveFence, joins today’s ‘#AIinBusiness’ podcast to discuss how enterprises can proactively test their AI systems to identify vulnerabilities before bad actors exploit them. From the origins of red teaming in military strategy to its modern applications in generative AI, he explains why organizations must move beyond compliance to build truly resilient AI systems. Different industries face unique AI security challenges. Financial services, healthcare, and consumer platforms all require tailored red teaming strategies to address evolving risks. Tomer explores how companies can align security with business goals, ensuring AI remains both effective and responsible. He also examines the growing trend of enterprises leveraging external expertise to simulate real-world threats, helping organizations stay ahead of adversaries. As AI adoption accelerates, businesses must strike the right balance between internal security capabilities and external partnerships. Red teaming is no longer optional—it is a necessary safeguard for AI integrity. Listen to the full discussion for insights into AI risk management, governance, and the future of AI security. #AI #MachineLearning #CyberSecurity #AIinBusiness #DataPrivacy #RiskManagement #GenerativeAI #AIAdoption #EnterpriseAI #SecurityStrategy #TrustAndSafety #ResponsibleAI #TechGovernance #AICompliance #BusinessLeadership

𝗧𝗵𝗲 𝗿𝗲𝘄𝗮𝗿𝗱 𝗼𝗳 𝗔𝗴𝗲𝗻𝘁𝗶𝗰 𝗔𝗜 𝗶𝘀 𝘄𝗼𝗿𝘁𝗵 𝘁𝗵𝗲 𝗿𝗶𝘀𝗸 – 𝘄𝗵𝗲𝗻 𝘆𝗼𝘂’𝗿𝗲 𝗽𝗿𝗲𝗽𝗮𝗿𝗲𝗱. As AI evolves, so do its challenges. #Agentic #AI, capable of autonomous decision-making and real-time interaction, brings unparalleled efficiency but also introduces some serious security risks. A proactive approach to mitigate these risks is crucial. Get the full breakdown in our latest #whitepaper: 𝗠𝗶𝘁𝗶𝗴𝗮𝘁𝗶𝗻𝗴 𝘁𝗵𝗲 𝗥𝗶𝘀𝗸𝘀 𝗼𝗳 𝗔𝗴𝗲𝗻𝘁𝗶𝗰 𝗔𝗜: 𝗔 𝗚𝘂𝗶𝗱𝗲 𝘁𝗼 𝗦𝗮𝗳𝗲 𝗗𝗲𝗽𝗹𝗼𝘆𝗺𝗲𝗻𝘁 𝗮𝗻𝗱 𝗨𝘀𝗲. 👉 Download now and deploy AI with confidence- https://lnkd.in/d-yrJZ6a

𝗚𝗲𝗻𝗲𝗿𝗮𝘁𝗶𝘃𝗲 𝗔𝗜 𝗶𝘀 𝗺𝗼𝘃𝗶𝗻𝗴 𝗳𝗮𝘀𝘁- 𝗵𝗼𝘄 𝗱𝗼 𝘆𝗼𝘂 𝗸𝗲𝗲𝗽 𝗶𝘁 𝘀𝗮𝗳𝗲 𝗶𝗻 𝗽𝗿𝗼𝗱𝘂𝗰𝘁𝗶𝗼𝗻? Join ActiveFence CTO, Iftach Orr, at NVIDIA GTC for a lightning talk on derisking generative AI in real-world applications. 📅 Monday, March 17 | 3 PM 💡 Session: The Real Last Mile- Derisking Generative AI in Production Taking generative AI from POC to production is just the beginning. Security, safety, and performance risks evolve in real time- so how do leading companies stay ahead ❓ In this session, you'll get practical insights, demos, and best practices from experts tackling LLM vulnerabilities, AI safety, and policy-driven safeguards. 🔥 𝗞𝗲𝘆 𝗧𝗮𝗸𝗲𝗮𝘄𝗮𝘆𝘀: ✅ Real-world examples of emerging Generative AI risks ✅ Tools & techniques for data protection, access control, and content safety ✅ How to build robust real-time evaluation pipelines ✅ Best practices for enterprises applying a holistic approach to AI risk If you're working with foundation models, AI agents, or mission-critical GenAI applications, this session is a must! 𝙒𝙞𝙡𝙡 𝙮𝙤𝙪 𝙗𝙚 𝙖𝙩 𝙉𝙑𝙄𝘿𝙄𝘼 𝙂𝙏𝘾? 𝘿𝙧𝙤𝙥 𝙖 𝙘𝙤𝙢𝙢𝙚𝙣𝙩 𝙤𝙧 𝘿𝙈 𝙞𝙛 𝙮𝙤𝙪’𝙙 𝙡𝙞𝙠𝙚 𝙩𝙤 𝙘𝙤𝙣𝙣𝙚𝙘𝙩! https://lnkd.in/dJxHN-H8 #GTC25 #GenerativeAI #AISafety #LLMSecurity #ActiveFence

𝗔𝗰𝘁𝗶𝘃𝗲𝗙𝗲𝗻𝗰𝗲 𝗶𝘀 𝗵𝗲𝗮𝗱𝗶𝗻𝗴 𝘁𝗼 𝗚𝗗𝗖 𝟮𝟬𝟮𝟱 𝘁𝗼 𝘁𝗮𝗰𝗸𝗹𝗲 𝘁𝗵𝗲 𝗯𝗶𝗴𝗴𝗲𝘀𝘁 𝘁𝗵𝗿𝗲𝗮𝘁𝘀 𝗶𝗻 𝗴𝗮𝗺𝗶𝗻𝗴. 𝗔𝗿𝗲 𝘆𝗼𝘂 𝗿𝗲𝗮𝗱𝘆? From AI-generated cheating and in-game toxicity to fraud and account takeovers, #gaming platforms are under attack like never before. At ActiveFence, we work with industry leaders to detect, prevent, and combat emerging threats- before they ruin the player experience. 𝗛𝗲𝗿𝗲’𝘀 𝘄𝗵𝗲𝗿𝗲 𝘆𝗼𝘂 𝗰𝗮𝗻 𝗷𝗼𝗶𝗻 𝘁𝗵𝗲 𝗰𝗼𝗻𝘃𝗲𝗿𝘀𝗮𝘁𝗶𝗼𝗻 𝗮𝘁 𝗚𝗗𝗖: ⭐ 𝗖𝗼𝗺𝗺𝘂𝗻𝗶𝘁𝘆 𝗖𝗹𝘂𝗯𝗵𝗼𝘂𝘀𝗲- We’re diving into the latest #trustandsafety challenges facing the gaming world. Join us to discuss how platforms can stay ahead of bad actors. 🥂 𝗚𝗮𝗺𝗶𝗻𝗴 𝗦𝗮𝗳𝗲𝘁𝘆 𝗖𝗼𝗮𝗹𝗶𝘁𝗶𝗼𝗻 𝗛𝗮𝗽𝗽𝘆 𝗛𝗼𝘂𝗿- Connect with fellow trust & safety leaders over drinks and discuss how we’re shaping a safer gaming future together. 👋 𝗠𝗲𝗲𝘁 𝘁𝗵𝗲 𝗔𝗰𝘁𝗶𝘃𝗲𝗙𝗲𝗻𝗰𝗲 𝗧𝗲𝗮𝗺- We’re on a mission to help gaming platforms fight back against evolving threats. Let’s chat- DM us to schedule a meeting! https://lnkd.in/dWYUJGeM See you at GDC! #GDC #GamingSafety #TrustAndSafety #ContentModeration #GameSecurity #ActiveFence

𝗕𝗲 𝗽𝗿𝗲𝗽𝗮𝗿𝗲𝗱 𝗳𝗼𝗿 𝘁𝗵𝗲 𝗿𝗶𝘀𝗸𝘀 𝗮𝘀 𝘆𝗼𝘂 𝗹𝗮𝘂𝗻𝗰𝗵 𝗚𝗲𝗻𝗲𝗿𝗮𝘁𝗶𝘃𝗲 𝗔𝗜 Every new AI system brings innovation and potential threats to safety and security. That’s why red teaming is a critical defense. ⚠️ But here’s the catch: without deep threat expertise, 𝗿𝗲𝗱 𝘁𝗲𝗮𝗺𝗶𝗻𝗴 efforts can fall short, leaving #AI systems exposed to vulnerabilities. 𝗪𝗵𝘆 𝗱𝗼𝗲𝘀 𝗲𝘅𝗽𝗲𝗿𝘁𝗶𝘀𝗲 𝗺𝗮𝘁𝘁𝗲𝗿? 🔴 Explore a few key reasons why expert-led red teaming is essential, then get more along with practical tips & best practices your teams can implement today. Don’t leave your Generative AI systems unprotected.  Learn more 👉 https://lnkd.in/ddwu2-Gi

🔓𝗨𝗻𝗹𝗼𝗰𝗸 𝘁𝗵𝗲 𝗦𝗲𝗰𝘂𝗿𝗶𝘁𝘆 𝗮𝗻𝗱 𝗦𝗮𝗳𝗲𝘁𝘆 𝗖𝗵𝗮𝗹𝗹𝗲𝗻𝗴𝗲𝘀 𝗼𝗳 𝗚𝗲𝗻𝗲𝗿𝗮𝘁𝗶𝘃𝗲 𝗔𝗜 Join ActiveFence and Amazon Web Services (AWS) experts for an exclusive discussion on building safer foundation models and AI applications- featuring Iftach Orr, Co-Founder & CTO of ActiveFence. 𝙒𝙝𝙖𝙩 𝙮𝙤𝙪'𝙡𝙡 𝙡𝙚𝙖𝙧𝙣: 💡The challenges and risks of not building safe Generative AI. 💡How to build GenAI applications with AI agents. 💡Essential guardrails & risk assessment strategies. This event is tailored for #AI developers, #security architects, and #tech leaders looking to enhance their 𝗔𝗜 𝘀𝗮𝗳𝗲𝘁𝘆 𝗮𝗻𝗱 𝘀𝗲𝗰𝘂𝗿𝗶𝘁𝘆 𝘀𝘁𝗿𝗮𝘁𝗲𝗴𝗶𝗲𝘀. 📅Date & Time: April 2nd, 2025 9:00 AM - 1:30 PM 📍Location: AWS Offices, Azrieli Sarona, Floor 28 𝗦𝗶𝗴𝗻 𝘂𝗽 𝗻𝗼𝘄: https://lnkd.in/dzvi3MZG Gili Nachum Keshet Brosh Yazan Khalaf Iftach Orr Itay Meller Gilad Sharabi

🔓𝗨𝗻𝗹𝗼𝗰𝗸 𝘁𝗵𝗲 𝗦𝗲𝗰𝘂𝗿𝗶𝘁𝘆 𝗮𝗻𝗱 𝗦𝗮𝗳𝗲𝘁𝘆 𝗖𝗵𝗮𝗹𝗹𝗲𝗻𝗴𝗲𝘀 𝗼𝗳 𝗚𝗲𝗻𝗲𝗿𝗮𝘁𝗶𝘃𝗲 𝗔𝗜 Join ActiveFence and Amazon Web Services (AWS) experts for an exclusive discussion on building safer foundation models and AI applications- featuring Iftach Orr, Co-Founder & CTO of ActiveFence. 𝙒𝙝𝙖𝙩 𝙮𝙤𝙪'𝙡𝙡 𝙡𝙚𝙖𝙧𝙣: 💡The challenges and risks of not building safe Generative AI. 💡How to build GenAI applications with AI agents. 💡Essential guardrails & risk assessment strategies. This event is tailored for #AI developers, #security architects, and #tech leaders looking to enhance their 𝗔𝗜 𝘀𝗮𝗳𝗲𝘁𝘆 𝗮𝗻𝗱 𝘀𝗲𝗰𝘂𝗿𝗶𝘁𝘆 𝘀𝘁𝗿𝗮𝘁𝗲𝗴𝗶𝗲𝘀. 📅Date & Time: April 2nd, 2025 9:00 AM - 1:30 PM 📍Location: AWS Offices, Azrieli Sarona, Floor 28 𝗦𝗶𝗴𝗻 𝘂𝗽 𝗻𝗼𝘄: https://lnkd.in/dzvi3MZG Gili Nachum Keshet Brosh Yazan Khalaf Iftach Orr Itay Meller Gilad Sharabi

𝗝𝗼𝗶𝗻 𝗔𝗰𝘁𝗶𝘃𝗲𝗙𝗲𝗻𝗰𝗲 𝗮𝘁 𝗡𝗩𝗜𝗗𝗜𝗔 𝗚𝗧𝗖! Excited for NVIDIA GTC? So are we! ActiveFence is bringing cutting-edge AI Security & Safety solutions to the Generative AI Pavilion- stop by our booth to see them in action!  𝗪𝗵𝗲𝗿𝗲 𝘁𝗼 𝗙𝗶𝗻𝗱 𝗨𝘀: 📍 Booth #3148 GenAI Pavilion | NVIDIA GTC 📆 March 18–21 🟢 Live Demos- See real-time risk assessment for Generative AI applications 🟢 Expert Insights- Chat with our team about securing LLMs and AI agents 🟢 Session Alert- Don’t miss our CTO Iftach Orr’s talk on derisking GenAI (March 17, 3 PM) Come by to explore how enterprises are tackling AI safety at scale- or DM us to book a meeting 👉 https://lnkd.in/dJxHN-H8 #GTC25 #GenerativeAI #AISafety #LLMSecurity #ActiveFence