Alice (Formerly ActiveFence)

London's getting a visit from the Alice crew next week. 🇬🇧 We're sending some of our finest to Infosecurity Europe, where "zero trust" is both a security framework AND a general life philosophy. Mo, Guy, and Tomer will be on the floor, ready to talk AI security, GenAI guardrails, and why your AI stack probably shouldn't be left unsupervised. Come find us. 👉 https://lnkd.in/dNz9xSV6 #InfosecEurope #CyberSecurity #AISecurity #CISO #London

When your AI guardrails were configured, was anyone from your compliance team actually in the room? If the answer is “not really,” there’s probably a gap between what your AI allows and what your policies actually say. Out-of-the-box guardrails know many, many rules. They don’t know YOUR rules. That’s why we built guardrails trained on your policies 👇 https://lnkd.in/d8Rutpkv

Conspiracy narratives after political violence don't emerge from nowhere. They follow a playbook. Alice's analysis of four major US political violence incidents, Butler, the Charlie Kirk killing, the WHCD shooting, and the latest White House attack, shows the same templates deploying each time, amplified by the same networks, increasingly seeded with AI-generated content before verified facts exist. Three TTPs we identified: - Template narratives adapted to each incident (staged event, deep state, foreign attribution) - AI-generated content seeding narratives before verified information exists - Domestic and foreign actors with pre-built distribution infrastructure activating within hours The result: nearly 30% of Americans now believe at least one Trump assassination attempt was staged. At Alice (Formerly ActiveFence), our Emerging Trend Detection team monitors these patterns continuously, so platforms, media organisations, and LLM systems can get ahead of misinformation before it embeds in the mainstream narrative. Read the full analysis: https://lnkd.in/d6sdFC5p #Misinformation #Disinformation #AIAbuse #SyntheticMedia #InformationIntegrity #TrustAndSafety #PoliticalViolence #OSINT #ActiveFence #Alice

What if the one thing AI can't swallow whole… is 𝘤𝘶𝘳𝘪𝘰𝘴𝘪𝘵𝘺? In an era where people are building their own replacement agents — and quietly watching their jobs get automated out from under them — what's the one thing AI can't devour? AI is changing work as we know it. And we're talking about it here. Next week, Mo sits down with James Villarrubia, former Presidential Innovation Fellow at NASA - National Aeronautics and Space Administration and now CTO, to say it plainly: AI might be Employee of the Month when it comes to productivity, but as James puts it, it isn't curious enough - It can answer almost anything. But it doesn't 𝘸𝘰𝘯𝘥𝘦𝘳. It doesn't ask the next strange, sideways question that cracks a problem open. And that, he argues, is the human skill AI can't replace. So who thrives in this new world? The worker with all the answers… or the one still asking better questions? 🤔 𝐂𝐮𝐫𝐢𝐨𝐮𝐬𝐞𝐫 & 𝐂𝐮𝐫𝐢𝐨𝐮𝐬𝐞𝐫, indeed. Full episode drops next Wednesday - listen wherever you get your podcasts. 🎧 #CuriouserAndCuriouser #StayCurious #AILiteracy #FutureOfWork #Curiosity #NASA #AIAdoption #CriticalThinking #Innovation #Podcast Madison

Your AI chatbot is customer-facing. Your regulators are watching. Here are four outputs that would make both of you very uncomfortable, and what purpose-built guardrails catch before they ship. 👇 https://lnkd.in/d7iCRRhc

We're proud to be partnering with Open Cybersecurity — The Trust Builders™ to bring AI red teaming, runtime guardrails, and board-ready audit trails to LATAM. A big thank you to Lucas Galvão and team for making this happen.

By popular demand (and after roughly 847 conference conversations that started with “wait… what exactly is AI red teaming?” 😅) We asked our CPO Avi Golan to break it down properly. So here it is: AI Red Teaming 101 - and why “traditional” security approaches stop working once AI systems become agentic. Because when AI can reason, use tools, make decisions, and act autonomously… You’re no longer testing static software. You’re testing behavior. In this session, Avi walks through: → why the agentic attack surface is fundamentally different → how real-world AI failures actually happen → why benchmarks and pen tests alone aren’t enough → and why it increasingly takes AI to break AI If you’re building copilots, agents, or customer-facing GenAI systems - this is one of those “everyone on the team should watch this” talks. Public service at this point 🎥👇 https://lnkd.in/dY2UdApf #AISecurity #AIRedTeaming #AgenticAI #CyberSecurity #LLMSecurity #AISafety

New disease, same playbook. Alice's Emerging Trend Detection team has been mapping health misinformation narratives for years, and the data from the 2026 hantavirus outbreak confirms what they've been seeing build across multiple crises: the infrastructure constructed during Covid-19 doesn't disappear between outbreaks. It waits. 92% of Hantavirus narratives mapped to the same broad framing seen during COVID. For #TrustandSafety teams, the takeaway is practical: treating each outbreak as a standalone misinformation event means you're always responding late. This piece lays out why that needs to change - and what to look for instead. Great work by Anna Baiges Capella.|\] Worth reading and sharing with your teams 👇

If AI risk has already made its way into your boardroom conversations - come find us at Gartner Security & Risk Management Summit. We're heading to National Harbor, June 1–3, with the best crew 👉 Noam, Eyal, John, Gustavo, Richard, and Smit. Book time with us ahead of the summit 👇 https://lnkd.in/dNKgjFqM We'll be easy to spot. 👋 See you there!

Sir Gavin has been bragging about his The New York Times subscription for years. Now it has finally paid off. 🐇 Don't miss our CEO Noam Schwartz, on why stripping #AI #guardrails just got dangerously easy, and what that means for everyone building with AI right now. Grab your coffee. It's a good weekend read. ☕ NYT piece → [https://lnkd.in/e-ff4AAj] The research behind it → [https://lnkd.in/dhfgEe2P] #AISecuirty #LLMSafety #AITrust #GenAI