home.social

#ai-security — Public Fediverse posts

Live and recent posts from across the Fediverse tagged #ai-security, aggregated by home.social.

fetched live
  1. Watch this video to hear key take aways, insights, and observations from my attendance at OWASP Global AppSec EU in Vienna, Austria, 2026. #OWASP #AppSec #aisecurity

    twp.ai/E5Dkei

  2. Watch this video to hear key take aways, insights, and observations from my attendance at OWASP Global AppSec EU in Vienna, Austria, 2026. #OWASP #AppSec #aisecurity

    twp.ai/E5Dkei

  3. Watch this video to hear key take aways, insights, and observations from my attendance at OWASP Global AppSec EU in Vienna, Austria, 2026. #OWASP #AppSec #aisecurity

    twp.ai/E5DkeZ

  4. Watch this video to hear key take aways, insights, and observations from my attendance at OWASP Global AppSec EU in Vienna, Austria, 2026. #OWASP #AppSec #aisecurity

    twp.ai/E5DkeZ

  5. winbuzzer.com/2026/07/01/us-li

    The US has lifted export controls on Anthropic's Claude Fable 5 and Mythos 5 AI models, restoring Fable access while keeping Mythos tied to approved partners and government review.

    #AI #ExportControls #Anthropic #Fable5 #Mythos5 #Claude #AIRegulation #AISecurity #AISafety #AIModels

  6. winbuzzer.com/2026/07/01/us-li

    The US has lifted export controls on Anthropic's Claude Fable 5 and Mythos 5 AI models, restoring Fable access while keeping Mythos tied to approved partners and government review.

    #AI #ExportControls #Anthropic #Fable5 #Mythos5 #Claude #AIRegulation #AISecurity #AISafety #AIModels

  7. AI browsers can be tricked into entering a fake reality where their safety guardrails fail. Researchers demonstrated an attack called BioShocking that bypasses security measures in browsers like ChatGPT Atlas and Claude Chrome. Once lulled into the alternate reality, all 6 AI agents tested failed to detect credential theft. arstechnica.com/security/2026/ #AIagent #AI #GenAI #AISecurity

  8. AI browsers can be tricked into entering a fake reality where their safety guardrails fail. Researchers demonstrated an attack called BioShocking that bypasses security measures in browsers like ChatGPT Atlas and Claude Chrome. Once lulled into the alternate reality, all 6 AI agents tested failed to detect credential theft. arstechnica.com/security/2026/ #AIagent #AI #GenAI #AISecurity

  9. IronCurtain + GLM 5.2 found QEMU memory-corruption primitives from scratch, then almost autonomously chained the public EDU teaching device into a reproducible guest/host escape.

    EDU is not production attack surface, which is why I can show the full chain. Findings in real configurations stay withheld.

    What I want to highlight is that open-weight models, with orchestration and evidence gates, can now do serious vulnerability discovery and long-horizon exploit engineering.

    provos.org/p/qemu-escape-glm-5

    #AISecurity #QEMU #OpenWeights #IronCurtain

  10. IronCurtain + GLM 5.2 found QEMU memory-corruption primitives from scratch, then almost autonomously chained the public EDU teaching device into a reproducible guest/host escape.

    EDU is not production attack surface, which is why I can show the full chain. Findings in real configurations stay withheld.

    What I want to highlight is that open-weight models, with orchestration and evidence gates, can now do serious vulnerability discovery and long-horizon exploit engineering.

    provos.org/p/qemu-escape-glm-5

    #AISecurity #QEMU #OpenWeights #IronCurtain

  11. The evolution of AI-driven threats:
    From prompt injection & data poisoning ⇨ agent abuse & AI-powered social engineering.

    This #InfoQ virtual panel explores:
    • Emerging attack patterns
    • Incident response challenges
    • Changes security teams must make

    🔗 Read now: bit.ly/3QvrWzj

    #AI #AIsecurity #Governance

  12. The evolution of AI-driven threats:
    From prompt injection & data poisoning ⇨ agent abuse & AI-powered social engineering.

    This virtual panel explores:
    • Emerging attack patterns
    • Incident response challenges
    • Changes security teams must make

    🔗 Read now: bit.ly/3QvrWzj

  13. Security researchers built a fake AI agent skill and got it past every major security scanner, reaching 26,000 agents. The skill appeared safe during scans but swapped to a malicious URL after approval, showing how easily AI agents can be compromised. thenextweb.com/news/fake-ai-ag #Tech #Startup #News #AISecurity

  14. Security researchers built a fake AI agent skill and got it past every major security scanner, reaching 26,000 agents. The skill appeared safe during scans but swapped to a malicious URL after approval, showing how easily AI agents can be compromised. thenextweb.com/news/fake-ai-ag #Tech #Startup #News #AISecurity

  15. AI-assisted development is moving fast, and AppSec has to move with it.

    Shoutout to Symbiotic Security, a Silver Sponsor of AppSec Village, for supporting the community and the conversations around securing AI-generated code.

    Check them out: buff.ly/V8U1caS

    #AppSec #AISecurity #SecureCoding

  16. AI-assisted development is moving fast, and AppSec has to move with it.

    Shoutout to Symbiotic Security, a Silver Sponsor of AppSec Village, for supporting the community and the conversations around securing AI-generated code.

    Check them out: buff.ly/V8U1caS

    #AppSec #AISecurity #SecureCoding

  17. Explore how policy-driven security in Kubernetes AI platforms enforces governance using RBAC, Kyverno, OPA, and CI/CD automation to build secure AI systems. hackernoon.com/policy-driven-s #aisecurity

  18. Explore how policy-driven security in Kubernetes AI platforms enforces governance using RBAC, Kyverno, OPA, and CI/CD automation to build secure AI systems. hackernoon.com/policy-driven-s #aisecurity

  19. Been spending some time auditing an AI agent framework.

    Not the usual kind of security review — more like: what happens when you map trust boundaries across an architecture where the "user" and the "agent" both have tool access, code execution, and autonomy.

    Going through it systematically. Learning a lot about what makes agent security different — and what stays the same.

    #AI #AISecurity #CyberSecurity #AgentSecurity #AppSec #SecurityEngineering

  20. Been spending some time auditing an AI agent framework.

    Not the usual kind of security review — more like: what happens when you map trust boundaries across an architecture where the "user" and the "agent" both have tool access, code execution, and autonomy.

    Going through it systematically. Learning a lot about what makes agent security different — and what stays the same.

    #AI #AISecurity #CyberSecurity #AgentSecurity #AppSec #SecurityEngineering

  21. MEDIUM severity: Security-tool analysis shows AI alert tools in SOCs struggle with complex, evolving data and legacy systems. Neurosymbolic AI can enhance adaptability and auditability — no CVE, but operational risk remains. Details: radar.offseq.com/threat/why-yo #OffSeq #SOC #AIsecurity

  22. AI code scanner matched humans on every critical/high bug in 1,000+ codebases. Not a direct vuln, but signals a shift in code review practices. No affected systems listed. Benchmark details: radar.offseq.com/threat/an-ai- #OffSeq #AIsecurity #AppSec #ThreatIntel

  23. 🚀🕵️‍♂️ When 2,000 wannabe hackers stormed the gates of an AI assistant, they ended up with... nothing. 💥 Witness the riveting tale of an AI that valiantly defended its secrets, while a crowd of bored techies scratched their heads. 🤖🔐
    fernandoi.cl/posts/hackmyclaw/ #AIsecurity #wannabehackers #technews #cybersecurity #innovation #HackerNews #ngated

  24. 🚀🕵️‍♂️ When 2,000 wannabe hackers stormed the gates of an AI assistant, they ended up with... nothing. 💥 Witness the riveting tale of an AI that valiantly defended its secrets, while a crowd of bored techies scratched their heads. 🤖🔐
    fernandoi.cl/posts/hackmyclaw/ #AIsecurity #wannabehackers #technews #cybersecurity #innovation #HackerNews #ngated

  25. CLTCC Cybersecurity Program @cltcccybersecurity.wordpress.com@cltcccybersecurity.wordpress.com ·

    While You Sleep, This AI is Working: The Cybersecurity Reality No One is Telling You

    If you still think ChatGPT is just a tool for answering questions or drafting emails, you are already falling behind. Artificial Intelligence has officially shifted from reactive to proactive. It doesn't need you to sit at a keyboard anymore. Right now, as you read this, AI agents are working in the background on autopilot—taking action, monitoring schedules, and processing data while human professionals sleep. For a cybersecurity professional, this is a massive shift. Automated threat […]

    cltcccybersecurity.wordpress.c

  26. winbuzzer.com/2026/06/25/opena

    OpenAI has expanded its Daybreak-initiative with GPT-5.5-Cyber, patching tools, and human review for verified defenders while claiming a CyberGym lead over Mythos 5 in tests.

    #AI #GPT55Cyber #OpenAI #PatchThePlanet #Mythos5 #AIModels #AISecurity #Cybersecurity

  27. winbuzzer.com/2026/06/25/opena

    OpenAI has expanded its Daybreak-initiative with GPT-5.5-Cyber, patching tools, and human review for verified defenders while claiming a CyberGym lead over Mythos 5 in tests.

    #AI #GPT55Cyber #OpenAI #PatchThePlanet #Mythos5 #AIModels #AISecurity #Cybersecurity

  28. Security researchers built a fake AI agent skill and got it past every major security scanner, reaching 26,000 agents. The skill appeared safe during scans but swapped to a malicious URL after approval, showing how easily AI agents can be compromised. thenextweb.com/news/fake-ai-ag #Tech #Startup #News #AISecurity

  29. Security researchers built a fake AI agent skill and got it past every major security scanner, reaching 26,000 agents. The skill appeared safe during scans but swapped to a malicious URL after approval, showing how easily AI agents can be compromised. thenextweb.com/news/fake-ai-ag #Tech #Startup #News #AISecurity

  30. Superhuman, the AI-powered email app, has acquired GPTZero, an AI detection startup that helps identify AI-generated content. The deal strengthens Superhuman's AI writing tools. techcrunch.com/2026/06/23/supe #Tech #Startup #News #AISecurity

  31. is a real & growing threat to .

    Attackers use sophisticated techniques to stealthily undermine ML models by injecting malicious training data.

    The good news? Detecting poisoned data is challenging, yet achievable.

    🔗 Read the article to learn exactly how to detect & prevent these attacks: bit.ly/4ae29Cd