AI 網路犯罪七大運作原則:為何駭客追求規模化,而非全自動攻擊
AI 尚未催生全新的網路犯罪模式,卻正讓既有犯罪手法變得更快速、成本更低,也更容易規模化。 趨勢科技 TrendAI™ 最新研究歸納出網路犯罪者採用 AI 的七大原則,解析 AI 如何提升網路釣魚與詐騙活動的執行效率、加速被竊資料的分析與處理,並逐步應用於惡意程式開發。同時指出,AI 代理程式雖已被用於實際入侵,但其自主攻擊能力仍受到技術可靠性、基礎設施存取權及人為監督等條件限制。

AI 尚未催生全新的網路犯罪模式,卻正讓既有犯罪手法變得更快速、成本更低,也更容易規模化。 趨勢科技 TrendAI™ 最新研究歸納出網路犯罪者採用 AI 的七大原則,解析 AI 如何提升網路釣魚與詐騙活動的執行效率、加速被竊資料的分析與處理,並逐步應用於惡意程式開發。同時指出,AI 代理程式雖已被用於實際入侵,但其自主攻擊能力仍受到技術可靠性、基礎設施存取權及人為監督等條件限制。
法律事務所的資安風險,未必會隨勒索軟體攻擊結束而消失。趨勢科技 TrendAI™ 研究團隊發現,網路犯罪者會在地下市場反覆兜售法律事務所的系統存取權限與機密資料,部分交易甚至在資安事件公開前就已出現,事件結案後仍持續流通。研究顯示,公開的勒索軟體資料外洩紀錄,可能只是法律產業所面臨資安威脅的一部分。
AI 代理程式(AI Agent)不只是聊天機器人,更能自主規劃與執行任務、呼叫外部工具,甚至與其他 AI 代理程式協作。隨著自主性提高,企業也面臨提示詞注入攻擊(Prompt Injection)、權限過大及機密資訊外洩等資安風險。 趨勢科技 TrendAI™ 透過實際測試,探討如何結合 NVIDIA OpenShell 與 TrendAI Vision One™,運用最小權限原則及涵蓋預防、偵測與應變的多層安全控管機制,協助企業降低 AI 代理程式帶來的資安風險,強化安全治理。
Google Chrome 會在符合條件的桌面裝置上自動下載約 4.27 GB 的 Gemini Nano 語言模型,使瀏覽器成為裝置端 AI 執行環境。然而,在現行架構下,任何 HTTPS 網站都能在沒有使用者同意提示、使用狀態指示及速率限制的情況下,於背景呼叫該模型,衍生提示詞注入(Prompt Injection)、運算資源濫用(Parasitic Computing),以及實驗性工具呼叫功能所涉及的潛在 Return-to-Tool(RTT)攻擊風險。 本文深入解析這三類資安風險,並提供企業資安團隊與開發者具體的防護及治理建議。
Researchers discovered a WordPress core vulnerability that can allow attackers to gain control of web sites and execute arbitrary code.
Read ArticleGoogle’s Home smart gadget infrastructure stopped working worldwide. Google released a fix that can be automatically installed after a reboot.
Read ArticleThe ePrivacy Regulation (ePR) is a proposed piece of legislation that wants to secure electronic communications in the European Union. Recent drafts show that the regulation may actually affect the development of applications like WhatsApp and Skype, as well
Read ArticleResearchers reported that the hospitality industry remains a top target for hackers aiming to steal data or access accounts using botnets and stolen credentials.
Read ArticleCloud security has become a significant issue as organizations increasingly move their workload to the cloud. A misconfigured cloud can leave organizations vulnerable to attacks.
Read ArticleThe UK tax authority has collected the voiceprints of over 5 million people, and a report from a privacy group shows that the organization may not be compliant with the new data protection regulations.
Read ArticleAn old tactic is being used by cybercriminals to bypass Microsoft Office 365 (O365) filters for phishing emails. The technique, called ZeroFont, involves the manipulation of text font sizes to trick O365’s natural language processing.
Read ArticleFour vulnerabilities, including two critical ones, in Phoenix Contact switches used in industrial systems automation could be exploited for attacks. Affected companies are strongly advised to apply patches immediately.
Read ArticleResearchers found a new malware called MyloBot in the wild that features new attack and evasion techniques — as well as the ability to delete other existing malware in the infected system.
Read ArticleResearchers found seven vulnerabilities in Axis cameras can be used for cyberattacks if exploited. While only POCs have been shown, customers are strongly advised to patch their firmware immediately.
Read Article