1,Analysis on the development status of the United States
As the most developed country in the content industry, the United States has the following characteristics in content security and attack and defense strategy: first, in view of the increasingly severe international situation and the needs of domestic ideological security, it has strengthened the supervision of the Internet content industry, especially on discrimination and prejudice; Second, attach great importance to the application of artificial intelligence algorithms in content security. Google and other companies cooperate closely with the government, and the government puts forward relevant review requirements for the security of algorithms; Third, both houses of Congress attach great importance to the use of artificial intelligence algorithms for content fraud, and incorporate the issue of content fraud into the legislative process by holding hearings or proposing relevant bills.
In 2019, the U.S. government issued the American artificial intelligence initiative, which emphasizes the importance of artificial intelligence to the traditional security field and ensures the leading position of the United States through artificial intelligence to meet the challenges from “strategic competitors and foreign competitors”. In March 2021, the National Security Council for artificial intelligence (nscai) of the United States issued the final report: actively maintain the dominant position of the United States in the field of artificial intelligence. The report systematically discusses how the United States can win the initiative and maintain the global leadership in the era of fierce competition in artificial intelligence, and expounds in detail the action route of future reform of federal agencies. The Advanced Research Projects Agency (DARPA) of the U.S. Department of defense launched the “algorithm warfare interagency team program” (i.e. “Maven program”) in 2017 to integrate machine learning algorithms into intelligence collection. DARPA has also promoted other research related to AI content security, including media forensics, interpretable AI, etc. In terms of content security technology based on artificial intelligence, the United States is also in the absolute leading position in the world. The citation influence, number of patents, number of enterprises and financing scale of American artificial intelligence papers rank first in the world.
2,Analysis of development status of European countries
European countries pay more attention to ethics in artificial intelligence security. In April 2019, the European Commission issued the code of ethics for trusted AI, which puts forward the framework for realizing the whole life cycle of trusted AI. The framework proposes seven key elements to achieve trusted AI: human initiative and supervision, technical robustness and security, privacy and data management, transparency, diversity, non discrimination and fairness, social and environmental well-being, and accountability.
Privacy and data management are particularly emphasized in the guidelines, which require that privacy and data protection must run through the whole life cycle of artificial intelligence system. Algorithms and data used in AI systems should be more transparent and traceable. When AI system has a significant impact on human beings, artificial intelligence system can reasonably explain the decision-making process and avoid unfair discrimination. Britain, France, Germany, Russia and other European countries have issued programmatic documents on artificial intelligence security, which are related to the strategy and development plan of artificial intelligence content security.
3,Analysis of Japan’s development status
In Japan’s Internet content industry, the most developed industries are games, animation, music and other industries. Therefore, in the era of artificial intelligence, Japan’s content security strategy mainly aims at the ownership of intellectual property rights of artificial intelligence production content, and promotes the introduction of relevant specific bills. In December 2018, the Japanese Cabinet Office issued the report on the social principles of human centered artificial intelligence, which is the highest level policy document issued by Japan to promote the development of artificial intelligence so far. It shows the position of the Japanese government in developing artificial intelligence from a macro and ethical point of view.
In the part related to the security of artificial intelligence content, the report mainly emphasizes the protection of privacy and security. In the principle of privacy protection, artificial intelligence can infer its political position, economic status, interests and hobbies with high precision based on personal action and other data. In the principle of ensuring safety, the report emphasizes that artificial intelligence systems need to balance benefits and risks. The report recommends in-depth study of AI risks and risk reduction methods, and pay attention to the sustainability of AI use.
4,Analysis of development status in other countries
In addition to the United States, Europe, Japan, China and other AI developed countries and regions, many other countries have also launched the national strategy of AI security, and content security is a part of it. In June 2018, India issued the national strategy for artificial intelligence, which proposed using artificial intelligence to promote economic growth and enhance social inclusiveness, and tried to find an artificial intelligence development strategy suitable for developing countries. In November 2019, Israel revealed its national artificial intelligence plan, which planned the development of artificial intelligence, how to meet the needs and security of the government and the military. The main goal is to determine that the education system and academic institutions can provide enough artificial intelligence engineers to meet the human needs of the government, national defense, military and industry. In recent years, the Canadian government has increased its investment in artificial intelligence research and development, and is committed to forming an extremely rich artificial intelligence ecosystem, including several specialized research institutions and thousands of artificial intelligence researchers. Researchers have conducted extensive research on content security oriented multimedia analysis, interpretable artificial intelligence and other technologies.