Experimental PromptLock ransomware uses AI to encrypt, steal data

Threat researchers have identified the first AI-powered ransomware, called PromptLock, which uses Lua scripts to steal and encrypt data on Windows, macOS and Linux systems. According to ESET researchers, PromptLock is written in Golang and uses the Ollama API to access the gpt-oss:20b large language model, with the LLM hosted on a remote server to which the threat actor connects through a proxy tunnel.

The malware relies on hard-coded prompts to generate malicious Lua scripts dynamically, including for local filesystem enumeration, target files inspection, data exfiltration and file encryption, and it is described as having data destruction functionality that has not yet been implemented. For file encryption, PromptLock uses the SPECK 128-bit algorithm, and there is a hard-coded Bitcoin address linked to Satoshi Nakamoto.

ESET told BleepingComputer that PromptLock has not appeared in their telemetry and is likely a proof-of-concept or work in progress, with some signs suggesting it is not an active threat in the wild. The article notes that in July Ukraine’s CERT reported the discovery of LameHug, another AI-powered tool, and that LameHug is believed to be deployed by Russian hackers of the APT28 group, using API calls instead of PromptLock’s proxying. 27 August 2025.