China has a right to access globally available intellectual property on artificial intelligence (AI) by legal means. It also contributes to the global fund of knowledge on AI. In May 2014, China’s search engine giant, Baidu, recruited Hong Kong born Andrew Ng, the head of Google Brain, to head its AI research. Chinese researchers are among the world’s leaders in this field.
But if recent observations by Bill Gates and Stephen Hawking are right, that artificial intelligence presents a threat to society in general, is that threat any greater or less in the case of China? If it is greater, is there any merit in considering export controls on artificial intelligence software and systems to China? What is the most effective policy response to China’s acquisition of advanced AI technology?
China has constructed the largest and most automated system for surveillance of its citizens ever seen in human history. As argued in my book, Cyber Policy in China, when it comes to choosing between e-democracy and i-dictatorship, the Chinese government has opted for the latter. China is already using highly automated systems for taking down internet traffic that it deems offensive. It has constructed a national database of all of its citizens, and it is building key grid-by-grid locality surveillance maps, including residency data, for sensitive parts of the country (such as Beijing and Lhasa). While the surveillance task China has set itself is for now beyond its technological means, a rapid development and application of AI to political censorship and surveillance by China could shift the current balance of power between Chinese netizens and their government heavily in favor of the latter.Enjoying this article? Click here to subscribe for full access. Just $5 a month.
The more famous gainsayers of artificial intelligence have not documented their concerns in any detail. Hawking went so far as to suggest it might spell the end of the human race. It would be a brave futurist who would attempt an assessment. But think what AI will come to mean in China’s case as long as the government continues to improve the automation of its regime of i-dictatorship.
Moreover, we can see a deepening pattern of cooperation between U.S. firms and U.S. investors in applying advanced AI to China’s security state. Apart from various law suits by various plaintiffs against major U.S. technology companies for supporting the repression of human rights in China through provision of related information technology, we need to only look to a web post about the 2013 China Security Expo to begin to see the dire long term evolution of China’s security state, ably assisted by free enterprise of the West. This one post reveals collaboration between two New York Stock Exchange listed companies (one Chinese and one American) collaborating in application of AI in developing and selling surveillance technologies. The post reads: “the Intelligent Video Surveillance System is the perfect combination of artificial intelligence and computer vision.” And that is only this decade’s AI-supported surveillance technology.
Once upon a time, a classic U.S. response to this sort of problem would be to control exports of sensitive AI technology to China. Well the AI genie is out of the bottle, out of the lab and the country. It was never wholly American to begin with. It is now a global enterprise, and a free one at that.
As we in the advanced democracies debate the impact of heightened AI-assisted mass surveillance at home in the name of defeating terrorism (the metadata privacy issue), we need to begin to include consideration of the China case and, separately, the future of AI in our debates. There is now a global challenge of newly empowered surveillance states being able to more rapidly enhance their power because of our support for their commercial adaptation of AI-assisted surveillance systems.
Governments in advanced democracies and their citizens should be alarmed that the window of opportunity for defeating automated mass surveillance in police states is closing, and each advance in AI is setting the pace for that fading opportunity.
There are policy responses available. For example, concerned actors outside China could usefully step up their dialogue with officials and civil society activists on the ethics of the information age. The Ministry of Public Security in China and its related universities need to be targeted aggressively on issues of AI ethics. But that is just one line of policy. We need a comprehensive analysis of this problem and widespread, robust engagement across all aspects of this issue if we are to stop advances in AI making the Chinese surveillance state the successful i-dictatorship its leaders want it to be.