人工智能安全问题研究

Research on security issues of artificial intelligence

近年来,人工智能又一次迎来发展的热潮,在自然语言处理以及数字图像处理等多个领城中的表现都超过人类。它已經被应用于教育、医疗、金融、军事等领城,与此同时,人工智能引发的一系列安全问题逐渐显现出来。人工智能系统的安全问题主要分类2类,即模型与数据隐私以及模型安全。对于不同的人工智能系统,它们处理的问题以及实现的原理都不同,因此,一个棘手的问题就是如何让防御措施更具有普适性,以处理各种对人工智能系统的安全威胁。

In recent years,artificial intelligence has once again ushered in a wave of development,outperforming humans in many fields such as natural language processing and digital image processing.It has been applied in fields such as education,healthcare,finance,and military,while a series of security issues caused by artificial intelligence are gradually emerging.The security issues of artificial intelligence systems are mainly classified into two categories,namely model and data privacy,and model security.For different artificial intelligence systems,the problems they handle and the principles they implement are different.Therefore,a thorny issue is how to make defense measures more universal to deal with various security threats to artificial intelligence systems.