The United States and China have reportedly reached an agreement to extend the decades-long cooperative framework of the US-China Science and Technology Agreement. That recent move follows Biden Administration hints from earlier in the year about intentions to cooperate with China on artificial intelligence safety.
What could possibly go wrong trusting the instinct to cooperate with Beijing?
When it comes to critical technology, a whole lot.
In 2006, Beijing signed a technology cooperation agreement with the United States, putting the Westinghouse AP1000 at the core of the Chinese nuclear energy program. Over the next decade, as Westinghouse worked with China, Chinese government-tied entities stole proprietary technical, design, and personal data from the company. By 2020, China had developed its own, State-owned domestic Westinghouse competitor – with leaders saying it “made no sense” to rely on foreign technology.
The track record isn’t much better when it comes to cooperation with Beijing on international security and public goods.
The US should lead and be willing to work with worthwhile partners, while punishing and holding accountable untrustworthy ones. At this point in US policy toward China, cooperation should be earned, not assumed.
Since 2006, China has repeatedly insisted that it abides by UN Security Council sanctions resolutions on North Korea – including sanctions for which Beijing voted. In 2022, the G7, EU and three other countries asked China to stop facilitating the entry of sanctioned petroleum productions into North Korea, to no avail.
Over the past couple decades, cooperation with the Chinese Communist Party has consistently not only failed, but also generated direct blowback: The PRC has stolen technology to support its non-market economy; transferred that technology to military use; encroached on public goods, including public goods it has promised to protect; and violated human rights and international norms at home and abroad. This reality, and the accompanying risks of partnership with China, are increasingly recognized.
And yet.
China weaponizes international cooperation on public goods
Despite Beijing’s track record – and despite the enormous stakes – the United States, as well as its allies and partners, continue attempting to carve out spaces for cooperation, hoping that Beijing will follow through on its commitments in those fields. Public health and climate have long been at the top of the list. Now, AI safety has joined them.
The argument is that these domains are public goods where the PRC’s cooperation is essential. That is a compelling argument: In other fields and with other partners, international cooperation has been critical to protecting the global commons and to expediting collective progress. Cooperation in space, for example, has yielded advances in everything from medical equipment to household cleaning products. More tactical international cooperation, like the Mine Ban Treaty, helps to decrease personal harms and make even formerly conflict-ridden territories safe.
But the argument doesn’t hold up if the PRC acts in bad faith, weaponizing the rest of the world’s cooperation to its own competitive advantage. And that is precisely what has happened to date.
Over the past couple decades, cooperation with the Chinese Communist Party has consistently not only failed, but also generated direct blowback.
Public health offers a stark and painful reminder: The COVID-19 virus likely emerged from a Chinese lab funded by cooperative dollars provided by the US government. That well-intentioned, cooperative endeavor has killed over 1 million Americans (and counting). And, regardless of COVID’s origin, Beijing refused to share critical information on the virus in the earliest and most critical stages of its global spread, all while stockpiling PPE months before raising the alarm. Hardly the stuff of protecting the global commons.
Or take climate: Beijing has made endless COP pledges. Yet coal has continued its banner run in China with a record new slew of coal-powered plants firing and starting construction last year. That coal (as well as forced labor) have powered PRC dominance in the solar energy industry that threatens the rest of the world’s energy security.
The risks of cooperating with a bad-faith actor
Now, the US is extending a cooperative hand to Beijing in AI safety – a field that the Chinese Communist Party explicitly categorizes as competitive, military-relevant, and zero sum. Even in its conciliatory propaganda, Chinese commentaries underscore the competitive nature of its positioning in AI as a “formidable challenge” to the assumed leadership of the United States. Hoping for cooperation with the CCP on AI is a fool’s errand: All past precedent, and all PRC policy, suggest that Beijing will leverage any concessions, support, or partnership the US provides for its own interests, and against those of the international system – better to acquire outside technology, establish control over global norms and standards, and limit restrictions on its AI policy. This will directly undermine US security and prosperity while hindering the competitive forces of the US technology ecosystem.
Global AI norms should be decided by democratic, market actors…The US government should know better at this point than to close its eyes and hope that there’s value in playing nice with Beijing.
Here is what risks playing out: Beijing will trade its purported cooperation on AI safety for relaxed restrictions on its access to cutting-edge US technology and US data. Beijing will also help shape, and sign onto, agreements that lessen US competitiveness in AI – while itself breaching those, probably flagrantly, and then demanding greater concessions for future compliance. Throughout, Beijing will be leveraging non-market means to define global AI norms and standards in ways that advantage its commercial champions and permit its human rights abuses.
Global AI norms should be decided by democratic, market actors. Those norms should empower market-motivated players to innovate and commercialize – with the full faith and backing of US and allied regulators, at the ready to punish Beijing’s non-market behaviors. The US should lead and be willing to work with worthwhile partners, while punishing and holding accountable untrustworthy ones. At this point in US policy toward China, cooperation should be earned, not assumed.
Beijing will trade its purported cooperation on AI safety for relaxed restrictions on its access to cutting-edge US technology and US data.
In the stories of COVID-19, Westinghouse’s stolen reactor technology, and false climate promises, there is a recurring bad guy: The Chinese Communist Party. The US-led global order rests on the benefits and imperatives of international cooperation. That is good. But if the US foolishly extends a cooperative hand to the bad guy, the good will be spoiled, weaponized, turned against the world. And at a certain point, it will be the fault of the US, as well: Fool me twice. The US government should know better at this point than to close its eyes and hope that there’s value in playing nice with Beijing. Charting an open-eyed path on AI – one wise to Beijing’s false promises – offers a chance to demonstrate an alternative, and to set a precedent across other critical areas of science and technology.
– Emily de la Bruyère and Nathan Picarsic, Senior Fellows at the Foundation for Defense of Democracies and co-founders Horizon Advisory
(Photo by Igor Omilaev on Unsplash)
