“抵御能力”
用英语怎么说？

The threat of digital deception is also mounting, as late-night television hosts find themselves the subjects of sophisticated AI deepfakes. Synthesized video clips of hosts like Jimmy Kimmel and Jon Stewart are proliferating online. Experts warn that these parodies are remarkably easy to produce and highly disruptive due to the trusted status of late-night figures. On a more positive note, a comprehensive assessment of AI chatbots yielded surprisingly encouraging results regarding their resilience against foreign disinformation campaigns. When tested with deceptive narratives propagated by hostile state actors, popular AI models successfully debunked falsehoods far more consistently than traditional search engines. However, AI-generated summaries at the top of search results performed noticeably worse, frequently propagating unsubstantiated claims and highlighting the critical necessity of ongoing algorithmic refinement.

数字欺骗的威胁也在增加，深夜电视节目主持人发现自己成了高科技人工智能深度伪造的目标。吉米·坎摩尔和乔恩·斯图尔特等主持人的合成视频片段正在网上激增。专家警告说，由于深夜节目主持人的公信力，这些恶搞视频非常容易制作，且具有极大的破坏性。从好的方面来看，对人工智能聊天机器人进行的一项全面评估在抵御能力外国虚假信息宣传方面取得了令人鼓舞的效果。当使用敌对国家传播的欺骗性叙事进行测试时，流行的人工智能模型比传统搜索引擎更始终如一地成功戳穿了谎言。然而，搜索引擎顶部的 AI 生成摘要表现明显较差，经常传播未经证实的信息，这突显了持续进行算法改进的紧迫性。