home.social

#llmjacking — Public Fediverse posts

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

fetched live
  1. LLMjacking si evolve: server Ollama esposti diventano il cervello di uno strumento di hacking autonomo, catturato in sviluppo da Sysdig

    Il Sysdig Threat Research Team ha catturato un threat actor che utilizzava un server Ollama esposto su internet come motore di ragionamento per VAPT, un framework offensivo automatizzato a più stadi. La campagna documenta la convergenza tra LLMjacking e offensive AI tooling: il modello AI decide ogni passo dell'attacco, dall'identificazione del servizio all'escalation dei privilegi.

    insicurezzadigitale.com/llmjac

  2. LLMjacking si evolve: server Ollama esposti diventano il cervello di uno strumento di hacking autonomo, catturato in sviluppo da Sysdig

    Il Sysdig Threat Research Team ha catturato un threat actor che utilizzava un server Ollama esposto su internet come motore di ragionamento per VAPT, un framework offensivo automatizzato a più stadi. La campagna documenta la convergenza tra LLMjacking e offensive AI tooling: il modello AI decide ogni passo dell'attacco, dall'identificazione del servizio all'escalation dei privilegi.

    insicurezzadigitale.com/llmjac

  3. 📢⚠️🤖 Operation Bizarre Bazaar has logged 35,000 attacks on exposed AI systems, stealing compute power and reselling access via the underground platform Silver Inc.

    Read: hackread.com/operation-bizarre

    #CyberSecurity #AI #LLM #BizarreBazaar #LLMjacking

  4. 📢⚠️🤖 Operation Bizarre Bazaar has logged 35,000 attacks on exposed AI systems, stealing compute power and reselling access via the underground platform Silver Inc.

    Read: hackread.com/operation-bizarre

    #CyberSecurity #AI #LLM #BizarreBazaar #LLMjacking

  5. 🚨 Microsoft Discupts Storm-2139 for LLMjacking & Azure AI Exploitation - Hackers hijacked AI models using stolen API keys.

    🔗 hackread.com/microsoft-storm-2

    #CyberSecurity #AI #LLMjacking #Microsoft

  6. 🚨 Microsoft Discupts Storm-2139 for LLMjacking & Azure AI Exploitation - Hackers hijacked AI models using stolen API keys.

    🔗 hackread.com/microsoft-storm-2

    #CyberSecurity #AI #LLMjacking #Microsoft

  7. 🚨 LLMjacking attacks are on the rise, now targeting DeepSeek and others! Attackers exploit stolen API keys, racking up massive cloud costs - one ORP instance generated $50K in 4.5 days.

    Read: hackread.com/hackers-monetize-

    #CyberSecurity #AI #LLMjacking #DeepSeek

  8. 🚨 LLMjacking attacks are on the rise, now targeting DeepSeek and others! Attackers exploit stolen API keys, racking up massive cloud costs - one ORP instance generated $50K in 4.5 days.

    Read: hackread.com/hackers-monetize-

    #CyberSecurity #AI #LLMjacking #DeepSeek

  9. 🚨 Speaking today at #NHICon Cybersecurity Conference! aembit.io/nhicon/ 🚨

    Excited to share that Crystal Morin and I will be presenting a session on LLMjacking Exposed: How Attackers Hijack AI Models, uncovering an emerging threat identified by the Sysdig Threat Research Team.

    🔍 What to expect:
    - An in-depth look at how attackers exploit large language models (LLMs) to enhance malicious activities.
    - A live prompt injection demonstration, showcasing how an LLM can be manipulated into executing malicious code.

    🗓 Join us live today, Jan 28th at 12:35 PM PST to explore this cutting-edge research!

    📢 Don’t miss out—register here: accelevents.com/e/nhicon

    #AI #Cybersecurity #LLMjacking

  10. 🚨 Speaking today at #NHICon Cybersecurity Conference! aembit.io/nhicon/ 🚨

    Excited to share that Crystal Morin and I will be presenting a session on LLMjacking Exposed: How Attackers Hijack AI Models, uncovering an emerging threat identified by the Sysdig Threat Research Team.

    🔍 What to expect:
    - An in-depth look at how attackers exploit large language models (LLMs) to enhance malicious activities.
    - A live prompt injection demonstration, showcasing how an LLM can be manipulated into executing malicious code.

    🗓 Join us live today, Jan 28th at 12:35 PM PST to explore this cutting-edge research!

    📢 Don’t miss out—register here: accelevents.com/e/nhicon

    #AI #Cybersecurity #LLMjacking

  11. @weirdwriter

    I'm shocked, shocked I say that computationally expensive autocorr-wrong is a massive security liability as well.

    Anyway here's some # that seem like they might be relevant.

    #infosec #AI #LLM #LLMjacking

  12. @weirdwriter

    I'm shocked, shocked I say that computationally expensive autocorr-wrong is a massive security liability as well.

    Anyway here's some # that seem like they might be relevant.

    #infosec #AI #LLM #LLMjacking

  13. Novel LLMjacking Attacks Target Cloud-Based AI Models

    Enter #llmjacking. It was probably inevitable. Threat researchers detected bad actors using stolen credentials to target LLMs, with the eventual goal of selling the access to other hackers.

    securityboulevard.com/2024/05/

  14. Novel LLMjacking Attacks Target Cloud-Based AI Models

    Enter #llmjacking. It was probably inevitable. Threat researchers detected bad actors using stolen credentials to target LLMs, with the eventual goal of selling the access to other hackers.

    securityboulevard.com/2024/05/