Artificial intelligence, explained.
Learning robotics in Bangkok, Thailand (Ana Norman Bermudez/UN Women)
Framing China’s public support for AI as an advantage nudges Americans to surrender debate, with costs for democracy.
Subscribe to The Informer for monthly expert analysis, and to Events for advance notice of visiting world leaders and distinguished guests.
You may unsubscribe from Lowy Institute newsletters at any time. For information on our privacy practices and how to unsubscribe, see our Privacy Policy.
The AI “enthusiasm” gap is a dangerous misnomer The AI “enthusiasm” gap is a dangerous misnomer
China’s new strategic advantage is characterised as a chasm (Opens in new window) between broad public approval of AI in China and intensifying scepticism in the US. But what seems like an asset may be a misnomer and a cautionary tale.
Renowned French philosopher Michel Foucault argued that narratives shape what people believe and how they act on those beliefs. Theorising language’s influence on human behaviour, he argued (Opens in new window) that “power is tolerable only on condition that it mask a substantial part of itself”. Foucault reminds us to question what phrases like Chinese “enthusiasm” and American “doom” direct us to think and do.
Recent policy articles (Opens in new window) and reports (Opens in new window) have sized up the AI arsenals of China and the US. They argue that while the US has a clear lead in specialised computer chips (Opens in new window), China maintains an advantage in public optimism. According to a KPMG survey (Opens in new window), 69% of the Chinese population believes that AI’s benefits outweigh its risks, compared with 35% of Americans. These findings reflect broad public concern in the US, which has ignited debates (Opens in new window) surrounding regulation ahead of next month’s midterm elections.
The KPMG study and broader “enthusiasm” framing also presume the Chinese public’s wilful embrace of technology, neglecting to account for the state’s centralised (Opens in new window), authoritarian (Opens in new window) integration of AI into society noted by researchers. While AI is embedded in daily routines and services in China from delivery robots (Opens in new window) to smart cars (Opens in new window), those who have spoken out against the externalities of digital tools have historically been silenced (Opens in new window) and arrested (Opens in new window). The state’s tight grip on technology has most recently involved barring (Opens in new window) AI executives from leaving the country who sold their start-up company Manus to Meta for $2 billion.
AI may erode freedom of expression in the US if left to the prerogative of the Trump administration.
China additionally offers a dystopian preview of how AI can further curtail and pre-empt public expression. The government currently uses (Opens in new window) AI to scan and censor swathes of textual and visual data, while Chinese companies are developing technology that would allow the state to reshape (Opens in new window) public opinion through personalised propaganda and predict (Opens in new window) which individuals will become future dissidents.
The word “gap” implies that there is something to fill, that the US is behind and Americans should hasten towards unflagging support. This presumes that the threat of China outpacing the US would inflict consequences far greater than those resulting from a lack of public debate. Such a cost–benefit analysis is questionable. The proliferation of low-cost Chinese models, including those released by DeepSeek and Z.ai (Opens in new window), is costing (Opens in new window) US companies market share as well as exporting (Opens in new window) political biases (Opens in new window) and supplying (Opens in new window) surveillance tools to autocracies.
However, while China’s access to public data has afforded (Opens in new window) it a scaling edge, it is the open-weight design that has made its models freely available, supporting its proliferation. Several approaches could allow US companies greater data access while preserving democracy and privacy. They include synthetic data generation (Opens in new window), which creates information from algorithms and simulation models that mimic (Opens in new window) real-world scenarios. A lack of domestic debate and regulation would not offset the iterative advantage of China’s open models. It could unravel American democracy from the inside.
AI may erode freedom of expression in the US if left to the prerogative of the Trump administration, which has repeatedly (Opens in new window) violated the constitutional protection of free speech (Opens in new window). Such a threat makes public debate surrounding AI a meta-safeguard for the very right to debate.
The Trump administration may also use digital tools to manipulate public opinion in elections, a tactic it employed – through human trolls (Opens in new window) – that predates the widespread and incisive competencies of AI. The risk of AI undermining free elections further underscores a need for public debate, which has animated both sides (Opens in new window) of the aisle.
Describing Chinese support for AI as the result of “optimism” and “enthusiasm” quietly counsel against such debate. Used to compare the US and China in what is characterised as a zero-sum race (Opens in new window) between the countries to dominate AI, these phrases suggest that the American public should surrender judgement calls surrounding regulatory guardrails. They imply that Americans’ concerns do not serve their self-interest. The irony of these descriptors lies in their exclusionary effects within the hawkish, binary framework of “us versus them”.
Given the Chinese government’s control over technology companies and their deployment of AI, “them” is clearly defined. “Us”, however, could mean tech executives who express societal aspirations for their R&D, politicians, civil society, or media in the United States. The appraisal of the American public’s concerns as a competitive disadvantage insinuates that the public should remove itself from decisions that will have significant political consequences.
Taking public pressure out of the equation may benefit AI executives and a White House that favours unilateral action. It is unlikely to benefit the American people, whose agency may be ephemeral as a result.
About the author
Sophie Berman
Sophie Berman is a researcher whose work focuses on the role of ethics in warfare, peace-making, and digital technologies. She holds an MPhil in International Relations from the University of Oxford.
The most-pressing world events explained by Lowy Institute experts and global contributors, in your inbox, every Thursday.
You may unsubscribe from The Interpreter at any time. For information on our privacy practices and how to unsubscribe, see our Privacy Policy.
