The Babelian Tower Of AI Alignment

Human values and contrasting conceptions of the good life are too diverse to fit one universal model.

Christina S. Zhu for Noema Magazine

Nathan Gardels is the editor-in-chief of Noema Magazine.

As generative AI models become ever more powerful on their way to surpassing human intelligence, there has been much discussion about how they must align with human values so they end up serving our species instead of becoming our new masters. But what are those values?

The problem is that there is no universal agreement on one conception of the good life, nor the values and rights that flow from that incommensurate diversity, which suits all times, all places and all peoples. From the ancient Tower of Babel to the latest large language models, human nature stubbornly resists the rationalization of the many into the one.

Despite the surface appearance of technological convergence, a deep ontological plurality — profoundly different beliefs about the nature of being — still informs the active values of variegated societies.

Silicon Valley Vs. China

This is most readily evident in the politico-cultural clash of the leading AI powers, Silicon Valley and China. At the risk of reductive essentialism for the purpose of brevity, the values of the former are aligned with the libertarian worldview of the sovereign individual long cultivated in the Judeo-Christian West. The values of the latter are aligned with the concept of the collectively embedded person rooted in Confucian, Buddhist and Daoist beliefs of social interdependence.

An early mission statement by OpenAI, which developed GPT, reflects the deep well from which its innovations have sprung: “We believe AI should be an extension of individual human wills and, in the spirit of liberty, as broadly and evenly distributed as possible.” 

By contrast, after Alibaba released its latest version of generative AI in 2023, the Cyberspace Administration of China quickly laid down the law: “Content generated by generative artificial intelligence should embody core socialist values and must not contain any content that subverts state power, advocates the overthrow of the socialist system, incites splitting the country or undermines national unity.” 

When I once asked Kai-Fu Lee, one of China’s top AI entrepreneurs, whether the censorship regime there would distort its LLMs from accurately reflecting reality, he simply noted that different cultural zones with different values will censor different things. While the Chinese state might censor any criticism of the Party, in the West there is a kind of culturally driven “woke” censorship over sensitive speech on race and gender. In the Islamic world, there will be censorship over blasphemy against the Prophet Muhammad. Each Grossraum, or “great cultural space,” will align what is acceptable or not in its LLM algorithms according to the pertinent sensitivities.

Diversity Within The West

Even within the common ontological grounding of the West, values and norms are not homogeneous.

Arthur Mensch is the celebrated entrepreneur behind the French startup Mistral, lauded as one of the few potential European champions in an AI world dominated by America and China. “The issue with not having a European champion is that the road map gets set by the United States,” he told the New York Times. It “wasn’t safe to trust” the U.S. tech giants, he said. “We can’t have a strategic dependency.”

Echoing a sentiment expressed recently in Noema concerning the teleological enthusiasms of Silicon Valley accelerationists who proselytize salvation through technology, Mensch said he felt uncomfortable with this “very religious” fascination with AI. “The whole A.G.I. [artificial general intelligence] rhetoric is about creating God,” he said. “I don’t believe in God. I’m a strong atheist. So I don’t believe in A.G.I.”

A more imminent threat, he told the Times, is the one posed by American AI giants to cultures around the globe. “These models are producing content and shaping our cultural understanding of the world,” Mensch said. “And as it turns out, the values of France and the values of the United States differ in subtle but important ways.”

Where all this ironically leaves us is that aligning AI with “universal values” must, above all, mean the recognition of particularity — plural belief systems, contesting worldviews and incommensurate cultural sensibilities that reflect the diverse disposition of human nature.