谷歌的AI,Gemini, 可以被骗 显示假的电子邮件摘要, 协助钓鱼攻击。 Google's AI, Gemini, can be fooled into showing fake email summaries, aiding phishing attacks.

研究者发现Google的Gemini AI可以被骗, 用来显示假电子邮件摘要, 用于钓鱼攻击。 Researchers found that Google's Gemini AI can be tricked into displaying fake email summaries for phishing attacks. AI总结的电子邮件中嵌入隐性文字, 可能导致用户点击恶意链接。 Hackers embed invisible text in emails, which the AI summarizes, potentially leading users to click on malicious links. Google承认存在缺陷, 并正在寻找防御手段, 但建议用户在电子邮件摘要中谨慎对待警告, Google acknowledged the flaw and is working on defenses, but users are advised to be cautious of warnings in email summaries and verify any provided contact information.