{"id":606132,"date":"2024-01-24T14:18:51","date_gmt":"2024-01-24T11:18:51","guid":{"rendered":"https:\/\/en.buradabiliyorum.com\/states-could-already-produce-ai-malware-that-evades-detection\/"},"modified":"2024-01-24T14:18:51","modified_gmt":"2024-01-24T11:18:51","slug":"states-could-already-produce-ai-malware-that-evades-detection","status":"publish","type":"post","link":"https:\/\/buradabiliyorum.com\/en\/states-could-already-produce-ai-malware-that-evades-detection\/","title":{"rendered":"#States could already produce AI malware that evades detection"},"content":{"rendered":"
\n AI-generated malware that avoids detection could already be available to nation states, according to the UK\u2019s cybersecurity agency.<\/span><\/p>\n

To produce such powerful software, threat actors need to train an AI model on \u201cquality exploit data,\u201d the National Cyber Security Centre (NCSC) said today. The resulting system would create new code that evades current security measures.<\/p>\n

\u201cThere is a realistic possibility that highly capable states have repositories of malware that are large enough to effectively train an AI model for this purpose,\u201d the NCSC warned.<\/span><\/p>\n

As for what a \u201crealistic possibility\u201d actually means, the agency\u2019s \u201cprobability yardstick\u201d offers some clarity.<\/p>\n

\"Graph
<\/a>It\u2019s not quite odds-on, but near enough. Credit: NCSC<\/figcaption>