30/10/2025
Artificial intelligence is already turning into a blackmailer. Cybersecurity specialists at ESET have discovered the first such threat—a family of ransomware programs that use artificial intelligence systems for local operations. Company representatives said that the ransomware is still in development but already has all the capabilities of traditional ransomware. The malware is called PromptLock, written in GoLang, and uses GPT-OSS:20b from OpenAI. ESET, the company that discovered the program, explained that PromptLock works on hard-coded prompts to generate Lua scripts “on the fly.” These scripts are then used to enumerate the file system, check files, steal data, and encrypt it. When extracting data, the program can not only encrypt it, but also destroy it. Versions of Promptlock exist for both Windows and Linux, and the Lua scripts they generate are compatible with various platforms. But there is good news. Promptlock attacks currently require several preconditions that are unlikely to be met in typical networks. For example, the success of a Promptlock attack depends on whether the victim has poor network segmentation and does not implement protective barriers or allow outbound traffic directed to LLM ports and protocols. In short, the malware is not yet a fully functional concept. But the developers will clearly improve their program soon, and we will soon face problems from AI. The Terminator will come sooner than we expect.
No one likes to be told the truth. We could get blocked. Subscribe to us on X and Telegram.
All Link: https://taplink.cc/wikileaksv2
Follow us on Telegram: https://t.me/wikileaks_v2
Follow us on X: https://x.com/WikiLeaksV2
Facebook Group: https://www.facebook.com/wikileaksversion2
