China on Wednesday launched its “Global Artificial Intelligence Governance Initiative” at the Belt and Road Initiative (BRI) forum in Beijing.
Although Chinese state media touted the A.I. initiative as an “open, inclusive, and fair approach” to A.I. development, the presence of international villains such as Russian leader Vladimir Putin and the Taliban at the BRI forum hinted at the true and much darker purpose of the Chinese Communist Party’s plans.
China’s state-run Global Times on Wednesday promoted the Global A.I. Governance Initiative as a “stark contrast to the U.S.’ restrictions and blockade aimed at preserving its hegemony.”
This was a reference to the U.S. and allied countries locking down sales of advanced computer technology to China last year, in recognition of the enormous security risk posed by the Chinese Communist Party’s malign activities, and its penchant for intellectual property theft. Tighter U.S. regulations on A.I. chips went into effect on Tuesday, infuriating Beijing.
“Even as China stressed fairness and non-discrimination in AI development so as to benefit all countries and regions, the U.S. intensified unilateral restrictions over normal global cooperation in A.I. and other technologies. Western countries have also been seeking to set rules for A.I., which experts say is mostly aimed at protecting their own interests,” the Global Times fulminated.
The thought of China’s gruesome authoritarian government gaining control of A.I. technology has been a nightmare plaguing the free world for most of the new century. Among the earliest concerns was that China’s lack of regard for privacy — for either its own subjects or foreign citizens — would allow it to compile gigantic databases to feed growing A.I. programs, leaping ahead of American and European companies bound by data privacy rules.
The Chinese Communist Party’s loose ethical standards quickly gave rise to horrific exercises in A.I.-driven surveillance, including research into A.I. facial recognition systems that would allow the tyrants of Beijing to track and monitor “troublesome” minority populations such as Tibetans and the Uyghur Muslims. A world in which China sets the standards for “A.I. governance” would be a world in which everyone but the political elite are monitored by relentless electronic eyes at every hour of the day.
The CIA and FBI have both expressed ongoing concerns about China’s influence on A.I. development and its abuse of the technology to drive massive espionage, surveillance, and influence campaigns.
FBI Director Christopher Wray said in January he was “deeply concerned” that China’s A.I. development is “not constrained by the rule of law.” Lakshmi Raman, the CIA’s director for artificial intelligence, warned only a few weeks ago that China’s A.I. programs are “growing every which way,” posing threats to both the security and stability of free nations.
The Department of Homeland Security (DHS) in September advised that adversaries like China are “seeking to undermine trust in our government institutions, social cohesion, and democratic processes are using AI to create more believable mis-, dis-, and malinformation campaigns.”
An example of that menace was uncovered by Microsoft in September: a Chinese-controlled network of social media accounts that used A.I. software to “mimic U.S. voters” and generate content that was “more eye-catching than the awkward visuals used in previous campaigns by Chinese nation-state actors.” The quality of the A.I.-generated propaganda improved noticeably as the system absorbed more information and refined its techniques.
China’s Global A.I. Governance Initiative talks about respecting “national sovereignty,” but examples such as the network discovered by Microsoft demonstrate that Beijing has no such respect for other countries. The launch announcement for the Chinese initiative even denounced “using A.I. technologies for the purposes of manipulating public opinion” — exactly what Microsoft just caught China using A.I. to do.
The Global Times ladled out Chinese propaganda buzzwords such as “fairness,” “openness,” “win-win cooperation,” and “non-discrimination,” but China clearly feels A.I. is a weapon it can wield carelessly because its own gigantic censorship apparatus and authoritarian Internet control makes it immune to the havoc it wreaks upon others. The most secretive, dishonest, and autocratic government on Earth feels very comfortable calling on everyone else to leave its digital doors unlocked.
There is some intriguing evidence that the Chinese Communist Party is wrong in that assessment. The Center for European Policy Analysis (CEPA) in August argued that A.I. could yet make the old dream of using Internet freedom to undermine authoritarian regimes a reality, reversing decades of dismal regression that made the Internet a better tool for tyrants than liberators.
CEPA noted that China hastily stuffed its homegrown A.I. chatbot genies back into their bottles because they kept coming up with unexpected and challenging answers to questions after being fed a steady diet of Communist propaganda. Chinese Internet giant Baidu just tentatively re-released its ERNIE chatbot on Tuesday after hauling it off to a digital re-education camp.
“Authoritarians rely on narrative homogeneity as a tool for control, viewing any deviation as a potential threat. Generative AI represents a threat to this narrative homogeneity. Authoritarians must contend with a plethora of news, media, and cultural productions that will challenge their accepted version of reality,” CEPA noted.
As CEPA explained, the early Internet was not too difficult for authoritarian regimes to control because search engines were “deterministic and easy-to-block.”
A.I. systems are driven more by input from consumers of information than content creators. They learn from both the questions they are asked, and the data they find — and information consumers vastly outnumber content creators. Social media unleashed a flood of information and emotions, which A.I. can focus into a data stream that authoritarians may find much more difficult to manage than the Internet of five years ago.
“With its ability to produce fully synthesized narratives, AI has the potential to expose users to new — and from the authoritarians’ perspective — dangerous narratives,” CEPA concluded.
The rulers of China will want to minimize that risk to their own power, so if the “Global A.I. Governance Initiative” is anything more than a temper tantrum at U.S. restrictions on advanced chip sales, it is a threat to strangle the new era of A.I.-driven provocation and rebellion in its crib.