PromptLock is the first documented AI-powered ransomware prototype, initially discovered by ESET researchers in August 2025 and later confirmed to be an academic proof-of-concept developed by researchers at the NYU Tandon School of Engineering. This malware distinguishes itself by its ability to dynamically generate its own malicious Lua scripts in real-time, leveraging a locally hosted Large Language Model (LLM) such as OpenAI's gpt-oss:20b via the Ollama API, enabling highly adaptive and polymorphic attacks. Its primary motivation is an academic demonstration to illustrate the potential dangers and novel attack methods of AI-orchestrated malware, with its design mimicking the goal of financial extortion through data encryption and exfiltration. The project is sometimes referred to as 'Ransomware 3.0' within academic contexts.

Paises afectados

URLs nuevas detectadas en IntelTracker