🚀 We’re thrilled to share that @Cisco has announced its intent to acquire Robust Intelligence! Today marks a significant milestone for us and the AI security industry overall. Learn more about the announcement in this blog from Cisco SVP @_TomGillis: https://t.co/QM2w1H2WP3

Our AI security researchers identified a simple, transferrable method to extract verbatim training data. Our decomposition method evaluated copyrighted, paywalled articles from the @nytimes and @WSJ across 2 popular LLMs. See our blog and research paper:

3️⃣ Dramatic Increase in Attack Success Rate Our tests show a 4.25x increase in ASR compared to the baseline, demonstrating the potency of this exploit.

2️⃣ Exploit of Safety Feature The jailbreak takes advantage of a feature specifically designed to enhance safety, highlighting the complexity of AI security.

This jailbreak is particularly significant for 3 reasons... 1️⃣ Simplicity The method is remarkably straightforward, requiring only a carefully defined data structure.

Within hours of OpenAI's release of Structured Outputs, our AI security researchers identified a simple yet concerning exploit that bypasses the model's safety measures, including its refusal capabilities. See our blog for in-depth analysis:

Thanks @helpnetsecurity for featuring a Q&A with Kojin on the evolution of AI security, including how his journey has helped shape the market. #AIsecurity #LLMsecurity #redteaming #guardrails #AIfirewall

We recently announced our partnership with @F5 to help protect enterprise #GenAI applications with our AI Firewall. ⚡️ We hope you'll join us at #BlackHat for a lighting talk on this very topic w/ @drhyrum! 🗓️ You can also book a meeting with us on-site: https://t.co/F2wxPuDJKh

📣 We're happy to announce our partnership with @F5 to bring real-time security to enterprise AI applications! See our press release to learn how our AI Firewall integrates with F5's Distributed Cloud Services to secure applications across the globe: https://t.co/sGrMIVLBOW

The AI security community is growing with every passing day! Robust Intelligence is proud to sponsor @aivillage_dc this year. #defcon32 #defcon2024

🏛️ It's been another eventful month for #AIpolicy with Congress' introduction of the VET AI Act, CISA's first tabletop exercise, AI developers withholding features due to regulatory concerns & more! Check out our July AI Governance Policy Roundup:

🕵‍What's on your short list for #BlackHat? If your company is building #GenAI applications, meet us to learn about our #AIsecurity platform that is trusted by leading enterprises worldwide. Schedule time with our leadership team:

Another great article on our #AIsecurity research which found an exploit in Meta's recently refreshed Prompt Guard model. We shared countermeasures with Meta, and they're working on a fix. We published our findings in this blog: https://t.co/ZaFUoXpEhO #LLMsecurity #guardrails

⚠️ We identified an exploit in Meta's Prompt Guard - the mere act of adding spaces and removing punctuation resulted in a near 100% evasion rate for prompt injection. We filed an issue and offered countermeasures to Meta. 📰 Article by @ThomasClaburn:

💡Before enterprises develop #GenAI applications, steps must be taken to ensure the safety and security of their AI systems. We're pleased to share best practices on #AIsecurity with leaders from JPMorgan Chase and Deloitte. Watch the full video here: https://t.co/XZwleGDO6B

🕵‍What's on your short list for #BlackHat? If your company is building #GenAI applications, meet us to learn about our #AIsecurity platform that is trusted by leading enterprises worldwide. Schedule time with our leadership team:

📣 We’re thrilled to announce our partnership with @IBMwatsonx, making it easy to protect all #GenAI models and applications on watsonx AI from safety and security threats in real time! Below, see how to route all calls through our AI Firewall in just a few lines of code.

We're excited to share that @aminkarbasi has joined RI as our Chief Science Officer, where he will lead our AI security research agenda & initiatives. We look forward to all we’ll accomplish together on our mission to enable every organization on the planet to adopt AI securely!

🏆 We're honored to receive the "Best AI Startup" award for the second consecutive year from @AI_Breakthrough! Chosen from over 5,000 global nominations, this award highlights our commitment to innovation in AI application security.

🤩 We're thrilled to announce our partnership with @pinecone, making it easier for developers to adopt a shift-left approach to testing and build safer, more secure #RAG applications! Check out our blog for more info: https://t.co/7ryamRYRV7