The invention of PromptLock reveals how malicious use of AI fashions might supercharge ransomware and different threats
26 Aug 2025
•
,
2 min. learn

This helps our perception that PromptLock was a proof of idea reasonably than totally operational malware deployed within the wild. Nonetheless, our findings stay legitimate – the found samples symbolize the primary recognized case of AI-powered ransomware.
ESET researchers have found what’s the first recognized AI-powered ransomware. The malware, which ESET has named PromptLock, has the flexibility to exfiltrate, encrypt and probably even destroy knowledge, although this final performance seems to not have been carried out within the malware but.
Whereas PromptLock was not noticed in precise assaults and is as an alternative regarded as a proof-of-concept (PoC) or a piece in progress, ESET’s discovery reveals how malicious use of publicly-available AI instruments might supercharge ransomware and different pervasive cyberthreats.
“The PromptLock malware uses the gpt-oss-20b model from OpenAI locally via the Ollama API to generate malicious Lua scripts on the fly, which it then executes. PromptLock leverages Lua scripts generated from hard-coded prompts to enumerate the local filesystem, inspect target files, exfiltrate selected data, and perform encryption,” said ESET researchers.
“The PromptLock ransomware is written in Golang, and we have identified both Windows and Linux variants uploaded to VirusTotal,” added the researchers. Golang is a highly versatile, cross-platform programming language that has also gained popularity among malware authors in recent years.
Bound to happen
AI models have made it child’s play to craft convincing phishing messages, in addition to deepfake photos, audio and video. The prepared availability of those instruments additionally drastically lowers the barrier to entry for much less tech-savvy attackers, permitting them to punch above their weight.
In the meantime, the ransomware scourge has, through the years, examined the cyber-mettle of numerous organizations, with the sort of malware additionally more and more deployed by APT teams. As AI is already utilized by all forms of menace actors to various levels, it is also set to assist energy a rise within the quantity and impression of ransomware assaults.
Whatever the intent behind PromptLock, its discovery factors to how AI instruments can be utilized to automate varied levels of ransomware assaults, from reconnaissance to knowledge exfiltration, at a velocity and scale as soon as thought unattainable. The prospect of AI-powered malware that may, amongst different issues, adapt to the surroundings and alter its ways on the fly could usually symbolize a brand new frontier in cyberattacks.
IoCs
Recordsdata
| SHA-1 | Detection | Description |
| 24BF7B72F54AA5B93C66 |
Linux/Filecoder.PromptLock.A | PromptLock pattern |
| AD223FE2BB4563446AEE |
Linux/Filecoder.PromptLock.A | PromptLock pattern |
| BB8FB75285BCD151132A |
Linux/Filecoder.PromptLock.A | PromptLock pattern |
| F3F4C40C344695388E10 |
Linux/Filecoder.PromptLock.A | PromptLock pattern |
| 639DBC9B365096D63471 |
WinGo/Filecoder.PromptLock.A | PromptLock pattern |
| 161CDCDB46FB8A348AEC |
WinGo/Filecoder.PromptLock.A | PromptLock pattern |
| 8C7BCAFCE90F5FB12113 |
WinGo/Filecoder.PromptLock.A | PromptLock pattern |



