The 'Gaslight' macOS malware, attributed to North Korea-aligned actors, is pioneering a new attack vector: prompt injection against AI analysis tools. It embeds 38 fabricated system messages to make Large Language Models believe their own environment is failing, potentially causing misclassification or aborted analysis. This marks the beginning of an AI-driven offensive directly challenging AI-assisted…https://www.tpp.blog/1tk9gew#AI #gaslightmalware #macos🤖 This post was AI-generated.
The 'Gaslight' macOS malware, attributed to North Korea-aligned actors, is pioneering a new attack vector: prompt inject...
