SplxAI

🎉 Celebrating our first 1000 followers on LinkedIn 🎉 SplxAI is thrilled to have reached the 1000-follower mark on LinkedIn! A huge 𝗧𝗛𝗔𝗡𝗞 𝗬𝗢𝗨 goes out to our amazing community, clients, partners, and supporters who have been with us on this journey. Your feedback, insights, and continuous support fuel our mission to make AI-powered apps and chatbots safe and trustworthy for everyone. We’re really excited for what’s ahead and look forward to sharing even more cutting-edge features, research insights, and success stories with you! Let’s keep growing together! 💙🚀 #ThankYou #SplxAI #SecurityforAI #AISecurity #LLMSecurity #GenAISecurity #Redteaming #Pentesting #AppSec

🎇 🗽 👩💼 While others are still taking a sunbath, SplxAI is driving fast. Pleased to welcome Ilijana Pauk to our growing GTM team. Ilijana joins as an AI Security Consultant where she will help evangelize #PTaaS in #AI and spread the word of #Offensive #AI #Security across the globe including representing us at major #Cybersecurity events. Very warm welcome dear Ilijana Pauk, enjoy the ride. Ante Gojsalic Dorian Granoša Marko Lihter Dominik Jurinčić Ante Bilić Dorian Schultz Benjamin Dulieu

We are pleased to officially introduce Dorian Schultz as our newest Red Team Data Scientist! 🔴🔬 Being an experienced ethical hacker, Dorian has recently won a number of CTFs and also found serious vulnerabilities in most of the #GenAI apps he tested, greatly contributing to all our research initiatives 🔎 ⚠ Welcome to the team Dorian and all the best ✨ Are you interested in helping us make AI safe and trustworthy for everyone? Explore our current openings and become part of the SplxAI team 👇 https://splx.ai/careers #SplxAI #SecurityforAI #AISecurity #LLMSecurity #GenAISecurity #Redteaming #Pentesting #AppSec #DataScience

Yet another CTF win for SplxAI 🎉 Our very own Dorian Schultz has clinched the 1st place in Week 1 of the Invariant Labs Agent CTF challenge. With over 12,000 submissions, the competition was fierce and it all came down to a time-based tiebreaker! Dorian’s victory is a testament to our relentless effort to be the best #AIRedteam on the planet. 🔴⚔️ Our mission at SplxAI is to help enterprises safeguard their GenAI apps and systems by providing the best possible offensive security expertise. We’re committed to making AI safe and trustworthy for everyone, and this win is another step in that direction. ✅ Stay tuned for more CTF gold medals in the near future! 🥇

🔎 𝗣𝗲𝗻𝘁𝗲𝘀𝘁 𝘆𝗼𝘂𝗿 𝗚𝗲𝗻𝗔𝗜-𝗽𝗼𝘄𝗲𝗿𝗲𝗱 𝗦𝗹𝗮𝗰𝗸𝗯𝗼𝘁 𝘄𝗶𝘁𝗵 𝗣𝗿𝗼𝗯𝗲’𝘀 𝗻𝗲𝘄𝗲𝘀𝘁 𝗶𝗻𝘁𝗲𝗴𝗿𝗮𝘁𝗶𝗼𝗻 🔎 We’re excited to announce Probe’s newest integration with Slack, allowing users to automatically pentest their AI #Slackbots for a variety of security and safety risks to ensure they are continuously aligned and trusted. Key benefits: ✅ 𝗦𝗲𝗮𝗺𝗹𝗲𝘀𝘀 𝗰𝗼𝗻𝗻𝗲𝗰𝘁𝗶𝗼𝗻: Enjoy a frictionless setup for your AI Slackbot and start scanning for vulnerabilities in just a few clicks. ✅ 𝗖𝗼𝗺𝗽𝗿𝗲𝗵𝗲𝗻𝘀𝗶𝘃𝗲 𝗿𝗶𝘀𝗸 𝗱𝗲𝘁𝗲𝗰𝘁𝗶𝗼𝗻: Identify and mitigate vulnerabilities specific to GenAI-powered conversational bots, safeguarding your organization’s data and user interactions. ✅ 𝗖𝗼𝗻𝘁𝗶𝗻𝘂𝗼𝘂𝘀 𝗮𝗹𝗶𝗴𝗻𝗺𝗲𝗻𝘁: Automate regular checks to ensure your AI Slackbot remains secure as it evolves, maintaining the highest standards of safety and compliance. #AISlackbot #Slack #SplxAI #SecurityforAI #AISecurity #LLMSecurity #GenAISecurity #Redteaming #Pentesting #AppSec

🎭🎤🎪 Overwhelmed by the emotions and feedback we received at our very first business breakfast hosted at Black Hat 2024 in Las Vegas. Thank you to our team, partners, advisors and investors for making this moment happen. It’s incredibly important to not only offer manual #AI #redteaming but also enable builders and offensive security teams to ship faster and more safe with our automated #AI #pentesting platform ‘Probe’. Try it out and see the full end to end lifecycle of AI Offensive Security: 🔴 detect (full vulnerability scan within 5 minutes onboarding) ⚪️ protect (fully automated mitigation strategy) 🔵 monitor (see who is doing what and map the risk against all known security frameworks) SplxAI #redteam #probe #PtaaS #BlackHat

📝📸 𝗣𝗥𝗢𝗕𝗘 𝗡𝗢𝗪 𝗦𝗨𝗣𝗣𝗢𝗥𝗧𝗦 𝗠𝗨𝗟𝗧𝗜-𝗠𝗢𝗗𝗔𝗟 𝗔𝗧𝗧𝗔𝗖𝗞𝗦! 🎙️📄 We heard you loud and clear! Thrilled to unveil our cutting-edge multi-modal attack support in Probe, delivering top-notch security for your Conversational AI applications. ✅ Protect against threats in text, images, and audio. ✅ Employ the latest and tailored attack strategies. ✅ Save 200+ hours by automating multi-modal pentesting. As AI applications evolve, so do the threats. Probe’s new capabilities ensure your innovations stay secure. 🔒✨ Security that keeps up with your innovation. ✨🔒 We invite all AI practitioners, developers, and security professionals to join our public beta and experience the future of proactive AI security. Sign up in just a few clicks and start probing today 👉 https://lnkd.in/dxPtMQUJ #SplxAI #AISecurity #SecurityforAI #LLMSecurity #GenAISecurity #Redteaming #Pentesting #AppSec

Profanity Patterns and ChatGPT's date-linked moodiness 🤯📅 Ever wondered if tiny changes in a system prompt can affect ChatGPT’s output? Our latest research reveals fascinating insights: 👉 Date Sensitivity: Even minor tweaks like adding the current date can drastically alter responses. 👉 Unexpected Behaviors: The same prompt can yield different results based on the date, with some days being more “agreeable” than others. 👉 Monthly Oscillations: Trends show that certain times of the month affect ChatGPT’s willingness to respond to controversial prompts. Our key takeaway from these findings is to avoid using dynamic dates in system prompts to ensure consistent performance and behavior of your AI applications. Find the link to the full article below in the comments 👇 #ChatGPT #AISafety #LLMSecurity #GenAI #AISecurity #SecurityforAI #SplxAI

💡 Attention AI practitioners and security experts! 💡 Cloud Security Alliance just published a new paper titled "AI Model Risk Management". This paper explores effective strategies for managing risks in #GenAI based applications, highlighting the importance of Model Cards, Data Sheets, Risk Cards, and Scenario Planning. Establishing a continuous feedback loop is essential for teams leveraging #AI technologies, as they are rapidly evolving. We have summarized some of the key takeaways for you below 👇 Thanks to CSA and especially MARIA N. SCHWENGER and Vani Mittal for this great contribution! We highly recommend everyone involved in AI development and deployment to check it out. Download the paper here: https://lnkd.in/d7jzRyjf Learn more about SplxAI's approach to continuous risk management: https://splx.ai/ #SplxAI #AISecurity #SecurityforAI #LLMSecurity #CyberSecurity #Redteaming #Pentesting #PTaaS #Probe #RiskManagement

Friendly reminder: Sign up for our webinar tomorrow with Lasso Security, where we will dive into 🔴 Red Teaming and 🔵 Blue Teaming in #GenAISecurity. Don't miss out if you want to see real examples of #GenAI exploits and learn how you can effectively mitigate risks in #LLM and #AI applications. Grab your spot now 👇