I think we need a lot more research on what China and other authoritarian anti-democratic states are doing with AI, and other such new tools for the control and subjugation of people, and, what are the implications inside and outside China. What’s the good research that has been done so far? including on the collusion by Western companies, and indeed similar developments in Western countries? Magnus Fiskesjö <firstname.lastname@example.org>
Source: The Economist (7/27/17)
Why China’s AI push is worrying: State-controlled corporations are developing powerful artificial intelligence
IMAGINE the perfect environment for developing artificial intelligence (AI). The ingredients would include masses of processing power, lots of computer-science boffins, a torrent of capital—and abundant data with which to train machines to recognise and respond to patterns. That environment might sound like a fair description of America, the current leader in the field. But in some respects it is truer still of China.
The country is rapidly building up its cloud-computing capacity. For sheer volume of research on AI, if not quality, Chinese academics surpass their American peers; AI-related patent submissions in China almost tripled between 2010 and 2014 compared with the previous five years. Chinese startups are attracting billions in venture capital. Above all, China has over 700m smartphone users, more than any other country. They are consuming digital services, using voice assistants, paying for stuff with a wave of their phones—and all the while generating vast quantities of data. That gives local firms such as Alibaba, Baidu and Tencent the opportunity to concoct best-in-class AI systems for everything from facial recognition to messaging bots. The government in Beijing is convinced of the potential. On July 20th it outlined a development strategy designed to make China the world’s leading AI power by 2030.
An AI boom in the world’s most populous place holds out enormous promise. No other country could generate such a volume of data to enable machines to learn patterns indicative of rare diseases, for example. The development of new technologies ought to happen faster, too. Because typing Chinese characters is fiddly, voice-recognition services are more popular than in the West; they should improve faster as a result. Systems to adjust traffic lights automatically in response to footage from roadside cameras are already being tested. According to the McKinsey Global Institute, a research arm of the consultancy, AI-driven automation could boost China’s GDP growth by more than a percentage point annually.
Yet the country’s AI plans also give cause for concern. One worry is that the benefits of Chinese breakthroughs will be muted by data protectionism. A cyber-security law that came into force in June requires foreign firms to store data they collect on Chinese customers within the country’s borders; outsiders cannot use Chinese data to offer services to third parties. It is not hard to imagine tit-for-tat constraints on Chinese firms. And if data cannot be pooled, the algorithms that run autonomous cars and other products may not be the most efficient.
A second area of unease is ethics and safety. In America, the technology giants of Silicon Valley have pledged to work together to make sure that any AI tools they develop are safe. They will look at techniques like “boxing”, in which AI agents are isolated from their environment so that any wayward behaviour does not have disastrous effects. All the leading AI researchers in the West are signatories to an open letter from 2015 calling for a ban on the creation of autonomous weapons. If it happens at all, the equivalent Chinese discussion about the limits of ethical AI research is far more opaque.
Chinese AI companies do have incentives to think about some of these issues: rogue AI would be a problem for the planet wherever it emerged. There is a self-interested case for the formulation of global safety standards, for example. But a third concern—that AI will be used principally to the benefit of China’s government—is a less tractable problem.
The new plan is open about AI’s value to the state. It envisages the use of the technology in everything from guided missiles to predictive policing. AI techniques are perfect for finding patterns in the massive amounts of data that Chinese censors must handle in order to maintain a grip on the citizenry. It is easy to imagine how the same data could boost the country’s nascent plans to create a “social-credit” system that scores people for their behaviour. Once perfected, these algorithms would interest autocratic regimes around the world. China’s tech firms are in no position to prevent the government in Beijing from taking advantage of such tools. Baidu, for example, has been appointed to lead a national laboratory for deep learning. Chinese AI will reflect the influence of the state.
Western firms and governments are no angels when it comes to data collection and espionage. But Western companies are at least engaged in an open debate about the ethical implications of AI; and intelligence agencies are constrained by democratic institutions. Neither is true of China. AI is a technology with the potential to change the lives of billions. If China ends up having most influence over its future, then the state, not citizens, may be the biggest beneficiary.
This article appeared in the Leaders section of the print edition under the headline “Code red”