当AI遇上AI:创新技术在识别和阻止金融诈骗中的前沿对抗  

How to Use Al Tools to Unravel New Types of Al Fraud

作　　者：韩裕 梁鹏 

机构地区：[1]云从科技集团股份有限公司

出　　处：《金融电子化》2024年第10期72-74,M0003,共4页Financial Computerizing

摘　　要：随着金融业务数字化进程加快,生物识别技术已经深度渗透到各个金融服务场景,包括但不限于手机银行、一体化柜面服务、自助银行终端、电话银行热线及信用卡应用程序等,通过人脸识别、声纹识别、移动端身份验证等多元化生物识别技术,有效提升了客户身份验证的安全级别。然而,在深度合成技术快速演进的当下,电信诈骗及AI换脸诈骗案件呈现出高发态势,传统基于活体检测的防御手段,在面对基于深度学习等先进算法生成的高度仿真视频及图像时,其识别效能面临严峻考验。The current generative Al technology is transforming the artificial intelligence industry while giving rise to new security risks.Large Models possess powerful learning and adaptive capabilities.However,this capability also means that Large Models could become new targets for hacker attacks and even be exploited by criminals as a new means to carry out malicious activities,posing severe cybersecurity threats.The focus of this issue's particular topic is financial security,which explores how to prevent Al fraud in the era of generative Al.

关 键 词：生物识别技术 身份验证 人脸识别 应用程序 深度学习 自助银行 电话银行 声纹识别 

分 类 号：F832[经济管理—金融学]