By simulating how AI can map systems, steal files, and craft ransom notes on its own, researchers warn that cheap, automated ransomware could soon outpace current defenses, making early preparation essential. https://lnkd.in/gQ2yVFr5
Ransomware AI: A new threat to cybersecurity
More Relevant Posts
-
Network & Security Engineer | Cisco, Palo Alto, Fortinet | Firewall & VPN Specialist | Securing Enterprise Infrastructures
Zero-Day AI Attacks: The Next Frontier in Network Security!! Recently, I came across some eye-opening coverage on how autonomous AI agents are evolving as a new threat vector, launching more personalized, hard-to-detect attacks rather than simply exploiting generic software bugs. https://lnkd.in/gzHYnu2K?
-
-
☠️ Anthropic’s latest Threat Intelligence Report reveals that a hacker with no coding skills orchestrated 17 sophisticated ransomware attacks using AI alone. The term “vibe-hacking” was coined to describe this new phenomenon where AI models like Claude Code act as operational threat agents; scanning networks, crafting malware, and even calculating ransom demands. One attacker compromised 17 organizations across healthcare and government sectors, all with AI as their silent partner. When I said that AI disrupts everything all at once, this is what I meant. There is no longer a barrier to entry in cybercrime. As AI becomes more capable, the line between tool and accomplice blurs.
-
When AI Meets Ransomware: Why gpt-oss:20b? Yesterday, researchers from ESET uncovered PromptLock – the first AI-powered ransomware, which, according to preliminary investigations, appears to still be in the POC stage. It leverages the new gpt-oss:20b model from OpenAI (via the Ollama API). The ransomware doesn’t come with a fixed payload—it generates Lua scripts on the fly while running. Everything is dynamic, adaptable, and flexible enough to tailor itself to the target in real-time. So why gpt-oss:20b? Why not gpt-oss:120b? After some thinking, it hit me: SnitchBench – a benchmark that measures how much models “snitch” to authorities or report crimes when presented with evidence. This model scored 0% Snitch Rate (where available). In other words: it simply doesn’t report. And what happened in practice? Attackers connected the dots: A powerful, free, open-source model Runs locally Doesn’t “snitch” The result: real-time generation of malicious scripts. In my view, the more interesting question isn’t just how this ransomware was built, but: Although ESET reported the model is not downloaded to the victim's machine, will AI monitoring become mandatory in every organization in the near future? How can we measure and assess the risk of AI models that “don’t report” malicious activity without stifling innovation? What do you think? #CyberSecurity #AI #Ransomware #EthicsInAI #SnitchBench #PromptLock
-
-
Vibe hacking is now a thing. “Agentic AI has been weaponized. AI models are now being used to perform sophisticated cyberattacks, not just advise on how to carry them out.” https://lnkd.in/dE7tmDQC
-
CRTP | Threat Intel | Darkweb | OSINT | Humint | Winner SIH 2022 | Threat Intelligence Analyst @Coralogix
AI and automation are no longer experiments—they are operational tradecraft Threat actors are rapidly evolving, leveraging AI and automation not only for technical sophistication, but also for novel extortion pressures. 🔎 In 2023, HYAS introduced BlackMamba—a polymorphic, AI-assisted malware that evaded leading EDRs by dynamically generating new code at runtime. 🧩 Just last week, ESET revealed PromptLock, an AI-powered ransomware concept so disruptive that yesterday LockBit group openly acknowledged it as an "Thank you for the idea". https://lnkd.in/gn4qAP9Z ⚙️ Today Huntress researchers exposed how adversaries are abusing automation platforms like Make.com to build scalable attack workflows for persistence and data theft https://lnkd.in/gjJTfhMp Now, Luna Lock a new ransomware group which was recently emerged has raised the stakes—not only encrypting and threatening to leak stolen data, but openly declaring that exfiltrated artwork will be submitted to AI companies for inclusion in training datasets if the ransom is not paid. this is probably just the beginning of what we’ll see in the coming days. #threatintel #darkweb #ransomware
-
-
Cybersecurity Leader with 20+ years of experience in enterprise environments. Certified Digital Forensics Analyst (GIAC), expert in Incident Response, Threat Intelligence, and Malware Analysis.
#APT groups are now wielding artificial intelligence as their newest weapon. Anthropic’s latest threat intelligence report reveals some eye-opening cases that show how AI isn’t just changing cybercrime, it’s revolutionizing it. #vibehacking https://lnkd.in/ecHeqdek
-
As an accomplished author and innovative entrepreneur, I am deeply passionate about exploring and advancing the synergy between Generative AI technologies and the transformative principles of Dr. Edwards Deming.
#AI AI-powered ransomware spotted: ESET researchers discovered what they described as the “first known AI-powered ransomware,” reportedly built using OpenAI technology. This reinforces concerns about criminals bypassing safeguards in major AI models like ChatGPT, Gemini, Llama, and Claude. https://lnkd.in/e9Vb72Xw
-
Cybercriminals are increasingly using generative AI tools to fuel their attacks, with new research finding instances of AI being used to develop ransomware. https://lnkd.in/e8ANvvUY