AI革新資安防護:主動防禦的未來

近期AI技術在資安領域的突破,標誌著資安防禦模式從「被動輔助」加速進化到「主動偵測與預防」,並推動團隊作業自動化、智慧化,特別在身份防護面向展現深遠影響。

首先,Google DeepMind與Project Zero團隊開發的AI代理「Big Sleep」於2025年7月成功預測並主動阻止針對SQLite的重大漏洞(CVE-2025-6965)攻擊。Big Sleep結合Google威脅情報與人類監督,在漏洞公開前即完成修補,是全球首例AI代理於實際環境中主動預防攻擊的案例,標誌資安防禦正式進入AI主動防禦新階段。此舉不僅有效防範潛在損害,亦推展至開源專案安全⁠⁠⁠⁠⁠⁠⁠⁠。

Google並於7月15日宣布多項AI資安新工具:Timesketch整合Sec-Gemini,賦予AI代理自動化初步取證分析能力,顯著縮短事件響應時間;FACADE則自2018年起於Google內部偵測內部威脅,採用對比學習(contrastive learning),無需歷史攻擊資料即可識別異常行為,提升未知威脅偵測效能⁠⁠⁠⁠。

身分防護方面,Okta與Palo Alto Networks宣布深度整合:不僅透過Prisma Access Browser與Okta Workforce Identity實現跨裝置安全存取,還結合Cortex XSIAM/XDR與Okta AI身分威脅防護,自動化高風險行為驗證與即時回應(如撤銷權限、隔離端點),協助企業建立零信任架構並降低資安複雜度⁠⁠⁠⁠。

在企業採用層面,非營利組織ISC²於7月16日發布「2025 AI Adoption Pulse Survey」指出,已有30%資安專業人士整合AI安全工具,42%正評估或測試中。AI主要應用於自動化重複性任務,讓團隊聚焦高複雜度威脅,28%受訪者認為AI為入門級人才創造新機會,44%認為對招聘影響有限。

然而,AI平台本身亦需強化資安。Meta AI於2025年1月修復一項隱私漏洞,該漏洞允許攻擊者透過推測提示編號存取其他用戶對話內容,凸顯生成式AI服務在存取控制上的潛在風險⁠⁠。

總結來看,AI已成為資安防禦新核心:一方面推動主動預測、即時攔截與自動化分析,另一方面在身分安全、企業應用與平台防護上發揮關鍵作用。展望未來,企業應同步規劃AI治理與風險控管策略,才能充分發揮AI在資安領域的潛力,迎戰日益複雜的威脅環境。

AI Security Defense Revolution! Latest Trends in July 2025

Hey! Security defense is undergoing dramatic changes! From passive “waiting to patch vulnerabilities" to active “preemptive strikes," and AI is the main character in this revolution!

The coolest thing is Google’s “Big Sleep" AI system, which discovered and patched a major SQLite vulnerability this July, giving hackers no chance at all! Google has also launched new tools like Timesketch and FACADE, making security analysis super fast.

In identity security, Okta and Palo Alto Networks have joined forces to create an ultra-strong protection system that can automatically handle risky behaviors and even directly isolate problematic devices.

Market research shows that 30% of security experts are already using AI tools, with another 42% currently testing them. But don’t forget, AI itself also has security issues, like when Meta AI had a vulnerability that allowed people to spy on others’ conversations.

In short, AI is making a splash in the security world, but companies should remember to simultaneously plan AI governance strategies to truly harness AI’s power and tackle increasingly complex security threats!