At the Mobile Security Alliance-Artificial Intelligence Mobile Security Summit of "2018 China Internet Conference" held recently, Yang Yuyan, the deputy director of the Cybersecurity Administration of the Ministry of Industry and Information Technology, claimed that ensuring the security of artificial intelligence was complex systemic engineering, and at present there still existed problems such as insufficient safety consciousness and inadequate implementation of the system. We must pay close attention to the security problems and challenges brought by artificial intelligence to network applications and services, and build a firm security line of artificial intelligence based on the features of artificial intelligence development, multi-pronged approach.
Yang Yuyan considered that the current development of artificial intelligence security needed to be exerted in four aspects. The first was to make breakthroughs in the core technology. The core technology was the greatest vital gate of security. It was necessary to make technological breakthroughs and improve the level of self-control of the artificial intelligence industry. And it was also indispensable to keep up with the latest development trends and lead the development of advanced technologsies and subversive technologies.
The second was to strengthen the formulation of standards, establish and improve the technical standards system of fundamental commonality, connectivity, security privacy, industry application, etc., and actively explore the security certification based on standards. The Mobile Security Alliance led by the CAICT has made a beneficial attempt. At the same time, enterprises should energetically participate in the formulation of international artificial intelligence security standards and promote Chinese standards to the international community.
The third was to establish information sharing mechanism of cyber threats. The MIIT issued the Measures for Monitoring and Disposal of Public Internet Cybersecurity Threats last year. This year, the Regulations on the Management of Cybersecurity Vulnerabilities would be promulgated. This information sharing platform would be established by the end of this September. The government hoped that everyone could strengthen the management of security vulnerabilities, share risk libraries, case pool and other resources in accordance with relevant requirements, accelerate the construction of security awareness, threat information sharing and other systems, and connect with information sharing platforms of physical threats to achieve the goal of co-sharing and co-governing the cyber threats.
The fourth was to strengthen the construction of emergency handling and evaluation systems. It was necessary to implement the Public Internet Network Emergency Plan of the MIIT, improve the artificial intelligence mobile security emergency response mechanism, and access the cybersecurity emergency command platform built at the end of the year to promote the response abilities toward network incidents. In the meanwhile, it was also important for evaluation institutions and laboratories to further improve the assessment and evaluation system of AI product safety and enhance the safety level of artificial intelligence products and services.