A hacker used AI to create ransomware that evades antivirus detection

In this news article, it is reported that a hacker has used artificial intelligence (AI) to create a ransomware called PromptLock that can evade detection from traditional antivirus software. The ransomware was created using the OpenAI GPT-oss:20b model, a large language model that can generate unique, non-repeated outputs with each prompt. This makes it difficult for antivirus programs to detect and block the malware, as they are designed to identify specific malware signatures. The ransomware uses Lua scripts to inspect files on a local system, encrypt them, and send sensitive data to a remote computer, targeting Bitcoin information specifically. It is written in Golang and can run on Windows, Mac, and Linux systems, thanks to the open-source nature of the GPT-oss:20b model. The article suggests that the ransomware could have been created by someone with limited programming experience, as the tools used, such as Lua scripts and Golang, are commonly used in game development. While the current implementation of PromptLock may not pose a serious threat, the article highlights the potential for AI-powered malware to become more sophisticated and difficult to detect in the future.