At the recent AI Action Summit in Paris, Vice President JD Vance declared that artificial intelligence must remain “free from ideological bias and … not be co-opted into a tool for authoritarian censorship.”
But China’s rise as a global AI power poses serious challenges to Vance’s bold vision. Chinese companies under the watchful eye of the Chinese Communist Party are quietly wiring digital authoritarianism into the Western tech stack.
Despite this threat, some Western governments and tech giants are pursuing ideological agendas in the form of “AI safety” that risk ceding moral and technological high ground to China in the AI race.
To reverse China’s digital creep and bolster Western technological supremacy, Europe and Big Tech should heed Vance’s admonition. Reject AI safetyism’s anti-Western social agendas and heavy-handed regulation in favor of “AI opportunity” propelled by American innovation and values.
China’s rise is a wake-up call.
Its recently unleashed DeepSeek R1 LLM appears to roughly match the performance of Open AI’s o1 model, one of America’s most advanced AIs.
But R1 comes with major strings attached. According to multiple investigative reports, the model is embedded with pro-Chinese Communist Party propaganda and censors sensitive topics such as Tiananmen Square and the government’s brutal oppression of the Uyghurs.
On top of that, R1 is capturing sensitive data and digitally fingerprinting millions of users—then feeding that information to a Chinese state controlled telecom provider, China Mobile.
Chinese tech and the authoritarian value system embedded within it are taking the Western tech stack by storm. Chinese AI companies are required by law to build their models to reflect CCP propaganda and aid the Chinese military to advance its totalitarian aims, both at home and abroad.
In mainland China, AI promises to enhance the CCP’s ability to surveil and control its citizens through an Orwellian social credit system. In Western nations, Chinese AI dominance could give the CCP a foothold to subtly influence the automated systems underlying everything from banking and HR, to search, shopping, and social media.
Winning the AI race is imperative for preserving freedom and defending the national sovereignty of Western nations against China’s digital creep.
But some Western governments and tech giants’ priorities are out of step, even contrary, to maintaining the moral and technological lead in AI.
At the Paris AI Action Summit, Vance emphasized the need for America and the West to lead AI development, in part, by preventing an over-zealous, heavy-handed approach to regulation.
Meanwhile, the summit’s European organizers, including France and the European Union, listed “making AI sustainable for people and [the] planet” as one of its main objectives.
That includes the Coalition for Sustainable AI, whose members, such as the Green Digital Action initiative, are prioritizing so-called green computing.
According to the organization’s website, its top goals include forcing tech companies to reduce carbon emissions by 45% by 2030 and driving “the adoption of international green standards” in tech.
But these policies risk starving Western AI companies of vital energy resources necessary to meet the substantial computational demands of their LLMs.
Consider that training OpenAI’s GPT-4 model alone consumed an estimated 50-gigawatt hours of electricity—2% of Sweden’s total data center energy capacity by some estimates. As “computational power is projected to double every 100 days,” AI models could increase global energy demand by between 26% and 36% annually.
We need vastly more carbon, not less, to power the computing infrastructure necessary to keep the West ahead of China in the AI race.
Beyond Europe’s climate alarmism, Western tech supremacy faces another threat from within: Ongoing attempts to bake DEI and “social justice” into AI and the tech stack under the banner of “safety.”
While the Trump administration is taking steps to expel ideological bias from AI, some European governments and Big Tech companies still embrace it.
For example, the Global Partnership on AI, which is comprised of the EU and member states, including the U.S., France, and Germany, commissioned a report in 2024 that explicitly calls for implementing “affirmative action across the AI ecosystem” to ensure that the “diverse perspectives” of “historically marginalized groups” are represented.
That’s code for designing AI models that censor their outputs to reflect political goals such as “fairness” and “equity.”
When AI is forced to do this, the result is often absurdities such as black Nazis, women popes, or comparisons of Elon Musk to Hitler.
Not only does this bias come at the cost of free speech, but it also detracts from model accuracy, meaning that so-called safer woke AI is less truthful and less useful to the end user.
This also works in China’s favor. Europe and Big Tech’s obsession with woke AI puts our tech firms more on the level of their Chinese counterparts, who must bias and censor models to please the CCP. China does this even though heavy-handed AI censorship may slow down development.
If Western AI firms, at the behest of Western governments, choose to censor their models to suit political sensibilities, they may very well be sacrificing a critical advantage over CCP-backed rivals. The freedom to build AI that tells the truth.
To beat China in the AI race, our governments and our companies must reject woke AI in all its forms.
The fate of Western democracy may well depend on it.
