Many scholars, analysts and other observers have suggested that resistance to innovation is the Achilles heel of authoritarian regimes. These governments may not keep up with the technological changes that help their opponents; they can also, by stifling rights, inhibit innovative economic activity and weaken the country’s long-term condition.
But a new study co-led by an MIT professor suggests something very different. In China, the research found, the government has increasingly deployed AI-powered facial recognition technology to crack down on dissent; has succeeded in limiting protest; and in the process, it has spurred the development of better AI-based facial recognition tools and other forms of software.
“What we found is that in regions of China where there is more unrest, this leads to greater government recruitment of facial recognition AI, subsequently by local government units such as departments of municipal police,” says MIT economist Martin Beraja, who is co. -author of a new paper detailing the findings.
What follows, as the paper notes, is that “AI innovation consolidates the regime, and the regime’s investment in AI for political control spurs further innovation at the border.”
Scholars call this state of affairs an “AI tocracy,” describing the connected cycle in which greater deployment of AI-driven technology stifles dissent while increasing the country’s capacity for innovation.
The open access document, also called “AI-tocracy”, appears in the August issue of the Quarterly Journal of Economics. Co-authors are Beraja, who is the Pentti Kouri Associate Professor of Economics for Professional Development at MIT; Andrew Kao, PhD candidate in economics at Harvard University; David Yang, professor of economics at Harvard; and Noam Yuchtman, professor of management at the London School of Economics.
To conduct the study, the scholars drew on several types of evidence spanning much of the past decade. To catalog cases of political unrest in China, they used data from the Global Database of Events, Language, and Tone (GDELT) Project, which records news worldwide. The team recorded 9,267 riot incidents between 2014 and 2020.
The researchers then examined records of nearly 3 million procurement contracts issued by the Chinese government between 2013 and 2019, based on a database maintained by China’s Ministry of Finance. They found that local government procurement of facial recognition AI services and complementary public safety tools (high-resolution video cameras) increased significantly during the quarter following an episode of public unrest in that area.
Given that Chinese government officials were clearly responding to public dissent activities by ramping up facial recognition technology, the researchers examined a follow-up question: Did this approach work to suppress dissent?
Scholars believe it did, although, as they note in the paper, they “cannot directly estimate the effect” of technology on political unrest. But as a way to get at that question, they studied the relationship between climate and political unrest in different areas of China. Certain weather conditions favor political unrest. But in China’s prefectures that had already invested heavily in facial recognition technology, these weather conditions are less conducive to unrest compared to prefectures that hadn’t made the same investments.
In doing so, the researchers also considered questions such as whether greater levels of relative wealth in some areas might have produced greater investment in AI-driven technologies, regardless of protest patterns. However, the scholars still came to the same conclusion: facial recognition technology was being deployed in response to earlier protests, and protest levels were then reduced.
“It suggests that technology is effective in calming riots,” says Beraja.
Finally, the research team studied the effects of increased demand for AI in China’s tech sector and found that the government’s increased use of facial recognition tools appears to be driving the technological sector of the country. For example, companies awarded contracts to acquire facial recognition technologies subsequently produce 49% more software products in the two years after being awarded the government contract than before.
“We look at whether this leads to greater innovation by facial recognition AI companies, and indeed it does,” says Beraja.
This data, from China’s Ministry of Industry and Information Technology, also indicates that AI-powered tools are not necessarily “excluding” other types of high-tech innovation.
Taken together, China’s case indicates how autocratic governments can reach a state of near-equilibrium in which their political power is enhanced, rather than reversed, when they take advantage of technological advances.
“In this age of AI, when technologies are not only growth-generating but also technologies of repression, they can be very useful” to authoritarian regimes, Beraja says.
The finding also relates to larger questions about forms of government and economic growth. A significant body of academic research shows that democratic institutions that grant rights generate greater economic growth over time, in part by creating better conditions for technological innovation. Beraja points out that the current study does not contradict previous findings, but by examining the effects of AI in use, it does identify a pathway through which authoritarian governments can generate more growth than they otherwise would have .
“This can lead to cases where more autocratic institutions develop alongside growth,” adds Beraja.
Other experts in the social applications of AI say the paper makes a valuable contribution to the field.
“This is an excellent and important paper that advances our understanding of the interplay between technology, economic success, and political power,” says Avi Goldfarb, the Rotman Chair in Artificial Intelligence and Health Care and professor of marketing at the University of Toronto’s Rotman School of Management. “The paper documents a positive feedback loop between the use of AI facial recognition technology to monitor local riot suppression in China and the development and training of AI models. This paper is pioneering research in AI and political economy. As AI becomes more widespread, I expect this area of research to grow in importance.”
For their part, scholars continue to work on aspects related to this issue. One of his forthcoming papers examines the extent to which China exports advanced facial recognition technologies around the world, highlighting one mechanism through which government repression could grow globally.
Research support was provided in part by the US National Science Foundation Graduate Research Fellowship Program; the Harvard Data Science Initiative; and the British Academy Global Chairs programme.
