Rohan Paul@rohanpaul_ai · X五眼联盟（澳大利亚、美国、英国、加拿大、新西兰情报共享联盟）罕见发布公开警告，称AI模型可能在数月内使针对政府和企业的毁灭性网络攻击变得极其容易。警告重点在于AI自动化专家级网络攻击任务：模型能读取代码、搜索漏洞、提出利用方案，并跨网络链式执行步骤，而不再需要安全专家手动逐一操作。
AI models capable of severe attacks on governments and businesses could arrive within months.
Five Eyes just issued a rare public warning， saying AI models could make devastating cyberattacks against governments and companies far easier within months.
Five Eyes is the intelligence-sharing alliance of Australia， the US， the UK， Canada， and New Zealand， built after World War II and still one of the closest spy and signals-intelligence partnerships in the world.
Their warning is mainly about automation of expert cyber jobs： a model can read code， search for weak spots， suggest an exploit， and chain steps across a network instead of waiting for one specialist to manually do each part.