当前位置:首页 > 报告详情

SANS360——下一场瘟疫.pdf

上传人: 可*** 编号:991765 2025-12-07 9页 1.10MB

1、The Next PlagueHow AI is Revolutionizing Malware DevelopmentSANS 360With Foster Nethercott./WHOAMI Foster Nethercott OSTACT13 on socials United States Marine Corps and Afghanistan Veteran I have 25 different industry certifications and an MSISE from SANS SEC535 Offensive AI Course Author I own Open

2、Source Tactician Suit EnthusiastnLowering Barriers to EntryHistorically,malware development required a fair amount of skill and technical expertise.Recent AI Advancements have all but removed any knowledge requirements to write unsophisticated malware.This is the next evolution of the“Script Kiddie”

3、nA Practical ExampleIn 2023,I wrote“The Evolution of the Digital Predator:Using AI to Evade Security Controls”The premise was simple.I acted as an individual with no coding knowledge,and wanted to test if I could get ChatGPT 4 to write undetectable malware.The subsequent program:Recorded all keystro

4、kes with time signaturesPerformed window monitoring,including URL tracking.Took ScreenshotsEstablished startup persistenceEstablished a C2 connection/exfiltration channelDisguised itself as Windows UpdaterA screenshot from a more recent program from SEC535nBut Most ImportantlyThe program is flagged

5、by 0 security vendors on Virus TotalnIncreasing Malware SophisticationAI Powered Poly and Metamorphic CodePerforming Real Time AdaptationsAI Powered Logic BombsDecoy Behaviors to fool Behavioral DetectionnNot Just TheoryIn 2018 IBM released research on Deep Locker,which acted as an AI powered Logic

6、BombFancy Bear(APT28)are alleged by Microsoft to be weaponizing AI for reconnaissance and malware development in the Ukraine war.Other major threat actors actively weaponizing AI for scripting and malware development include Charcoal Typhoo

word格式文档无特别注明外均可编辑修改,预览文件经过压缩,下载原文更清晰!
三个皮匠报告文库所有资源均是客户上传分享,仅供网友学习交流,未经上传用户书面授权,请勿作商用。
根据报告的内容,全文主要内容概括如下: - **AI革命化恶意软件开发**:AI技术使得编写恶意软件不再需要专业技能,降低了进入门槛。 - **AI恶意软件实例**:作者通过ChatGPT 4生成了一款未被发现的安全软件,展示了AI在恶意软件开发中的应用。 - **AI恶意软件特点**:包括多态和变形代码、实时适应、逻辑炸弹和欺骗行为。 - **AI恶意软件案例**:IBM的Deep Locker、Fancy Bear(APT28)在乌克兰战争中的AI武器化、HYAS Labs的BlackMamba等。 - **应对策略**:加强研究、使用可解释人工智能(XAI)、减少门控、增加威胁情报共享、投资AI防御工具。
"AI助长恶意软件?揭秘新威胁!" 恶意软件开发新纪元!" 恶意软件的未来走向?"
客服
商务合作
小程序
服务号
折叠