西方可以向全球南方的 AI 樂觀者學到甚麼

On a recent Sunday afternoon, I was invited to debate AI’s impact on the cultural sector alongside Dutch politicians and creative-industry leaders. I was the lone optimist, arguing that if fear and risk become our primary lens for governing artificial intelligence, we will design systems around what we want to prevent rather than what we want to use them for. Then I said something that was harder for the room to hear: The Global South is home to some of AI’s most enthusiastic and inventive users, and their experiments could help us rethink our own institutions. 

The U.S. is no different from Europe. A culture of fear encourages protectionism: We hunker down and romanticize legacy systems rather than ask what might replace them. Creative industries in the majority world have spent decades navigating piracy and weak enforcement, forcing them to think beyond rights-based policing about how creators actually receive value. Anghami, the biggest music app in the Middle East, for instance, pays musicians from their first stream, rather than only after a song reaches a minimum threshold. It has approached AI pragmatically, using it to curate, tag, and attribute creative work rather than simply blocking and banning.

In the West, fear of AI has helped turn pessimism into a cultural virtue. Would you publicly say you love AI? Your answer may say something about where you’re from. Say “yes” to a concerned audience in the U.S. or the Netherlands, and you risk being seen as naive, or complicit in extraction and exploitation. While the U.S. is building and deploying AI at a scale no other country matches, Americans are deeply anxious about where it is heading. Only 17% of U.S. adults think AI will have a positive impact on the country over the next 20 years, according to a 2025 Pew Research Center survey. 

In much of the Global South, the mood is markedly different. Stanford’s 2025 AI Index found that 83% of respondents in China, 80% in Indonesia, and 77% in Thailand said AI-powered products offer more benefits than drawbacks, compared with 39% in the U.S. and 36% in the Netherlands. 

So why the divide?

It’s all about access

Demographics are a part of the answer. Nearly 90% of the world’s young people live outside the West. They are not simply using AI, they are playing with it. Drive through Hyderabad or Bengaluru and you will see billboards for AI crash courses and Bollywood films promoted as AI-made. I have spent two decades talking with young people from the rural Himalayas to the slums of Hyderabad, and they often tell me some version of the same thing: They love technology. Their enthusiasm is not necessarily about the technology itself, it is about what AI might democratize: access to institutions and opportunities that have long excluded people with the least social or cultural capital.

A founder I met in Gurgaon told me she is anxious about AI too, but that it has allowed her to reach clients far beyond the English-speaking world, reinventing her business in a fraction of the time it once took. A civic activist in Chennai taught herself to code with AI and built a gender violence-alert app in multiple languages, at a moment when funding for such work is drying up and state scrutiny is rising. Having seen what deepfake abuse does to women, she sees AI as one of the few tools powerful enough to fight back. India consistently ranks among the most optimistic countries on AI even as anxiety about jobs rises.

It would be easy for the West to dismiss these stories as products of scarcity, shaped by conditions far removed from our own. That would be a mistake. Constraint has a way of forcing people to rewrite the rules of the game. At a technology festival in Athens this year, creatives told me they worry AI might kill short filmmaking. In China, AI-made micro-dramas have become a multibillion-dollar industry. One breakout film, Zombie Cleaner, was made for less than $500 by a part-time wedding photographer in rural Yunnan province, and went on to attract tens of millions of views overseas. The significance is not the novelty of AI-generated entertainment. It is that AI lowers the cost of entering a cultural market. People in smaller Chinese cities can tell stories in their own accents and languages rather than waiting for mainstream media to tell them what counts as a story.

Building for need, not market share

A few months ago, I spoke to employees at a multinational company about “inclusive AI.” I was asked to remove a slide on AI’s potential harms because it might discourage adoption. That exchange captures something important about AI adoption in the West. Corporations increasingly pressure employees and customers to use AI, often because adoption itself has become a measure of competitiveness. The question is how to maximize market return.

Elsewhere, the question is more often: What problem can this solve?

Where capital and resources are scarce, technologies are frequently built around concrete public needs. India, with a fraction of Europe’s resources, built India Stack: a public digital infrastructure for identity, payments, and data-sharing that companies can build on rather than infrastructure controlled by a handful of private platforms. Europe is now looking to India’s example as it develops EuroStack, in response to the growing need for digital sovereignty.

Innovation does not always come from having more. Sometimes it comes from having less room for failure. The countries with the least room for error are experimenting with AI not because they are blind to its risks, but because they cannot afford to wait for perfect systems to solve chronic problems. The West should pay closer attention.

The point is not to romanticize the Global South, but to recognize that we have a remarkably narrow vision of what AI can be for. We ask whether it will replace workers, destroy creativity, undermine democracy, or kill our planet. These are necessary questions. But what if we also asked what AI could help us rebuild? Young people around the world are using it for work, entertainment, romance, creativity, activism, and self-expression. They are not waiting for institutions to tell them what the technology is supposed to be for.

If the most powerful builders keep designing for market share while the world’s most resource-constrained builders design for survival and self-actualization, the templates that spread globally for governance, infrastructure, and who gets to participate may increasingly come from wherever necessity did the inventing. The West is warning the world about what AI might do to us. Perhaps it is time to ask what the rest of the world might teach us to do with it.