aifollow.news
返回 IT之家 科技新闻
IT之家 科技新闻· · 原发布时间 AI 评分40

华中科技大学与阿里巴巴共建 AI 安全研究中心,启动 9 个课题

AI 辅助摘要

华中科技大学与阿里巴巴集团签署合作协议,共建智能攻防与前沿安全联合研究中心。首批 9 个课题涵盖大模型安全攻防、智能体与终端安全等四个方向,其中两项将分别探索识别手机智能体操作中的画面伪造风险,以及通过密态计算保护跨应用数据协同中的隐私。

正文 · AI 翻译

IT Home reported on November 21 that Huazhong University of Science and Technology and Alibaba Group signed a cooperation agreement on November 18 to jointly establish a “Joint Research Center for Intelligent Attack and Defense and Frontier Security” focused on AI security, with an initial 9 research projects launched.

IT Home learned that the joint research center’s initial 9 projects cover 4 areas: offensive and defensive security for large models, agent and device security, intelligent systems and software security, and data collaboration and privacy protection. The following are cases officially made public:

When AI agents begin operating phones on users’ behalf and completing complex tasks, screen images could be forged, and the operation process could be subject to malicious manipulation. In response, the joint research center has established a project on “Adversarial Image Examples and Device Threat Assessment” to explore technologies for identifying and defending against the associated security risks.

When users switch between shopping, maps, food delivery and other apps, they want the convenience of data sharing but worry that their personal information could be exposed during collaboration. To address this, the “Privacy-Preserving Relationship Graph” project will explore technologies such as privacy-preserving computation to make data “usable but not visible, and collaborative without direct transmission,” so that convenience and privacy protection no longer have to be an either-or choice.

发现内容有误?提交纠错