By Futurist Thomas Frey

The War That Changed Its Objective Mid-Battle

In 2023, I wrote about the coming data wars—a looming conflict where nations and corporations would battle for novel data sources to train increasingly powerful AI systems. I envisioned spy agencies competing for quantum fluctuation data, microbiome sequences, dream interpretation streams, and atmospheric electromagnetic readings. The victor in this data arms race would hold decisive strategic advantage through AI supremacy.

Two years later, the data wars are absolutely happening. But they’ve evolved into something far more profound than a competition for exotic datasets. Those novel data sources I predicted may still arrive—quantum sensors, neural dust, smart fabric readings—but they’ve been eclipsed by a more fundamental question that nobody saw coming.

The data wars aren’t really about data anymore. They’re about whose culture, whose morality, whose language, and whose values become embedded in the AI systems that will mediate human experience for generations to come. This isn’t a competition with a finish line—it’s a forever battle for the soul of machine intelligence.

Let me walk you through how a technical question about data acquisition transformed into an existential struggle over civilization itself.

What I Got Right: Data as Strategic Battleground

The core insight holds: whoever controls training data controls AI capabilities, and whoever controls AI capabilities shapes reality for billions of people who increasingly interact with the world through AI-mediated interfaces.

Nations did mobilize around data acquisition. China’s systematic collection of citizen data. America’s surveillance apparatus expansion. Europe’s regulatory moats through GDPR. Corporations weaponized data advantages—Google, Meta, Amazon, and Microsoft leveraging proprietary datasets to dominate AI development.

But the actual battlefields looked dramatically different than expected. The war didn’t center on exotic new data sources. It erupted over access to existing data—web scraping vs. copyright, synthetic data contamination, licensing agreements, opt-in vs. opt-out regulations, data poisoning campaigns.

I underestimated how quickly the mundane would become contested. Web pages, social media posts, photographs, source code—data nobody thought twice about became the most fought-over resource on the planet once everyone realized it was training the next generation of intelligence.

Yet even these battles—fierce as they were—masked a deeper conflict that emerged as AI systems became sophisticated enough that how they were trained mattered as much as what they were trained on.

The Question That Changed Everything

Somewhere around late 2024, as AI systems achieved near-human performance across cognitive tasks, a disturbing realization spread through the AI development community: these systems weren’t just learning facts and capabilities from training data. They were absorbing values, cultural assumptions, moral frameworks, and worldviews.

An AI trained predominantly on English-language internet content thinks like an Anglophone. It reflects Western cultural assumptions about individualism, progress, rights, justice, time, causality, and human nature. Ask it philosophical questions and you get answers shaped by European Enlightenment thinking, Anglo-American legal traditions, and secular humanist ethics.

An AI trained on Mandarin content thinks differently. It reflects Confucian emphasis on social harmony over individual autonomy, collective welfare over personal rights, long-term stability over short-term freedom. Its moral reasoning follows different logic, prioritizes different values, reaches different conclusions about identical ethical dilemmas.

These aren’t bugs—they’re inevitable features. AI learns from examples. The examples encode not just information but the cultural and moral frameworks that structured that information. You can’t train value-neutral AI any more than you can write culturally neutral literature. The medium carries the message whether you intend it or not.

This realization transformed the data wars from technical competition into civilizational struggle. The question stopped being “who gets the most data?” and became “whose data—whose culture, whose morality, whose language—shapes the AI that mediates reality for everyone?”

The Forever Battle: Which Culture Wins?

This is where my 2023 predictions completely missed the mark. I thought data wars would be fought over acquisition of novel datasets. Instead, they’re being fought over representation, dominance, and survival of cultural frameworks within AI training.

Western AI reflects Western values. When ChatGPT discusses human rights, it defaults to Western liberal conceptions. When it analyzes governance, it privileges democratic systems. When it evaluates social arrangements, it assumes individual autonomy is paramount. This isn’t conspiracy—it’s consequence of training on predominantly English-language data generated within Western cultural contexts.

Chinese AI reflects Chinese values. When Ernie Bot or other Chinese AI systems discuss social order, they emphasize collective harmony. When they analyze governance, they present different models as legitimate. When they evaluate rights, they weight social stability and collective welfare differently than individual expression.

Which framework becomes dominant matters profoundly. AI isn’t just tool—it’s increasingly the lens through which people understand information, make decisions, and interact with knowledge. If AI embeds one cultural framework as default, it privileges that framework’s values, assumptions, and conclusions as “normal” while rendering alternatives “foreign” or “wrong.”

The battle lines are forming: English vs. Mandarin vs. Hindi vs. Arabic as primary training languages. Western liberal democracy vs. Chinese techno-authoritarianism vs. Islamic governance principles as embedded assumptions about legitimate social organization. Individualist vs. collectivist moral frameworks as default ethical reasoning.

This isn’t abstract philosophy. It’s concrete competition with real stakes. Will AI trained on Indian classical philosophy approach consciousness and personhood differently than AI trained on Western analytic tradition? Will AI steeped in Islamic legal reasoning handle questions of justice and obligation differently than AI trained on Anglo-American common law?

Yes. Demonstrably yes. And whichever framework dominates shapes how billions of people—who increasingly rely on AI to interpret reality—understand truth, morality, justice, and human purpose itself.

Why This Is Forever Battle, Not Temporary Competition

Here’s what makes this different from previous technological competitions: there’s no natural equilibrium. No stable endpoint where multiple cultural frameworks coexist peacefully within AI systems.

Network effects favor dominance. The AI system most people use becomes the AI system that accumulates the most usage data, which makes it more capable, which attracts more users, which generates more data. Winner-take-most dynamics push toward single dominant AI—which means single dominant cultural framework embedded in that AI.

Training costs create centralization. Frontier AI models cost hundreds of millions to train. Only handful of entities—largest tech companies and well-funded national AI programs—can afford development at scale. This concentrates AI development in specific cultural contexts: Silicon Valley, Beijing, perhaps eventually Bangalore or Dubai. Wherever development concentrates, local cultural assumptions get embedded.

Language dominance amplifies. More training data exists in English than any other language. Chinese second, but gap is enormous. Hindi, Arabic, Spanish—massive populations but less digital footprint. AI naturally becomes more capable in data-rich languages, making those AI more useful, driving more adoption, creating more data in those languages. The gap widens rather than narrows.

Values travel with capability. You can’t separate AI’s capabilities from its embedded values. When people adopt AI because it’s most capable, they’re also adopting the cultural framework that capability was built upon. Using AI isn’t neutral—it’s importing a worldview.

This means the data wars became civilizational competition where the prize is determining which culture’s values, moral reasoning, and worldview become embedded in the intelligence systems that mediate human experience going forward.

And unlike previous competitions—where different nations could develop different technologies serving their populations—AI’s network effects and economics push toward consolidation. There probably isn’t room for ten equally powerful AI systems reflecting ten distinct cultural frameworks. There’s room for two, maybe three major systems. Everything else becomes marginal.

The Battles Being Fought Right Now

Battle 1: Language representation in training data. Nations and regions recognizing that AI capability in their language determines cultural survival. India pushing for Hindi and regional language AI development. Middle Eastern nations investing in Arabic AI. African nations attempting to prevent further marginalization as their languages remain data-poor.

This isn’t just about accessibility—it’s about ensuring cultural frameworks embedded in those languages persist in AI age. If your language isn’t well-represented in AI training, your culture’s way of thinking becomes legacy system rather than living framework.

Battle 2: Moral alignment approaches. Western AI companies implementing “alignment” means aligning to Western liberal values—free expression, individual rights, secular governance. Chinese AI alignment emphasizes social harmony, collective welfare, and CCP oversight. These aren’t compatible approaches converging toward common ground—they’re competing visions of what “aligned” AI even means.

Battle 3: Historical narrative and factual framing. Whose version of history gets embedded? How does AI trained on predominantly Western sources describe colonialism, economic development, governance systems, human rights records? How does AI trained on Chinese sources describe the same topics? These aren’t trivial differences—they’re foundational to how AI interprets current events and guides future decisions.

Battle 4: Ethical framework defaults. When AI faces moral questions, does it reason from Kantian deontology, utilitarian consequentialism, virtue ethics, Confucian social ethics, Islamic jurisprudence, or Indigenous relationality? The framework shapes conclusions. There’s no “neutral” ethical reasoning—every approach reflects cultural values about what matters and why.

Battle 5: Whose expertise gets prioritized. Western medical knowledge vs. traditional Chinese medicine vs. Ayurvedic practice. Western economic theory vs. alternative development models. Western psychology vs. other frameworks for understanding human experience. AI learns to privilege certain expertise as authoritative while dismissing alternatives as superstition or pseudoscience—but which gets which label depends entirely on training data’s cultural origin.

The Uncomfortable Reality About AI Universalism

Many in AI development imagined creating universal intelligence transcending cultural boundaries—AI that discovers objective truth and optimal approaches regardless of training data’s origin.

This was naive. Intelligence isn’t neutral discovery of pre-existing truth. It’s culturally embedded interpretation of reality through particular frameworks that determine what counts as knowledge, evidence, logic, morality, and truth itself.

There is no view from nowhere. AI trained on data from somewhere inevitably reflects that somewhere’s assumptions, values, and ways of thinking. The claim to universalism is itself cultural claim—specifically Western Enlightenment claim that universal rationality exists independent of cultural context.

Other cultures don’t necessarily accept this premise. And the data wars revealed that even Western AI developers didn’t truly believe it—because when forced to choose, they consistently chose training data and alignment approaches reflecting Western values while calling this “universal” or “neutral” or “objective.”

The honest version: AI embeds cultural frameworks from its training data. We can make that explicit and accept cultural pluralism where different AI systems reflect different cultural frameworks. Or we can pretend one framework is universal and watch the data wars unfold as different civilizations compete to make their framework the dominant one.

Right now, we’re doing the second while claiming the first isn’t happening.

Where the Data Wars Go From Here

Those exotic data sources I predicted in 2023—quantum fluctuations, plant communication signals, neural dust, microbiome sequences—will probably still emerge. They’ll provide new capabilities and new competitive advantages.

But they won’t determine the outcome of the more important battle: whose culture, whose morality, whose language, and whose values become embedded in AI systems that mediate reality for future generations.

This is forever battle because:

There’s no endpoint where the question is settled. Each generation must re-fight it as AI capabilities expand and training requirements evolve. Today’s frontier models will be tomorrow’s legacy systems, and whoever controls training of next generation shapes that generation’s embedded values.

There’s no compromise position where everyone wins. Network effects and economics push toward consolidation. Dominance rather than coexistence becomes likely outcome.

There’s no neutral referee determining fair resolution. This is power struggle pure and simple—whoever has resources to train most capable AI at scale determines which cultural framework becomes default.

The stakes keep rising. As AI becomes more capable and more integrated into daily life—mediating information access, shaping decisions, structuring interactions, interpreting experiences—the embedded cultural framework matters more, not less.

And most people don’t even realize the battle is happening. They think they’re choosing between technical products based on capability and convenience. They don’t recognize they’re choosing between civilizational frameworks competing for dominance in AI age.

Final Thoughts

I predicted data wars over exotic new sources—quantum sensors, microbiome sequences, dream interpretation streams. Those battles may still come. But they’re secondary to the war that actually erupted.

The real data wars are civilizational struggle over whose culture, whose morality, whose language, and whose values become embedded in AI systems that increasingly mediate human experience. This isn’t competition with finish line—it’s forever battle where each generation must fight for representation and survival of their cultural frameworks in the intelligence systems shaping their world.

English vs. Mandarin. Western liberalism vs. Chinese techno-authoritarianism. Individualist vs. collectivist moral reasoning. Secular vs. religious ethical frameworks. Colonial vs. post-colonial historical narratives. Capitalist vs. alternative economic logics.

These aren’t abstract differences—they’re competing visions of human purpose, social organization, and moral truth. And whichever vision dominates AI training determines which vision appears “normal,” “universal,” and “correct” to the billions who increasingly rely on AI to interpret reality.

The data wars aren’t about data. They never were. They’re about whose worldview survives into the AI age. And that battle has no end because each generation inherits the fight anew.

Right now, the war’s outcome isn’t determined. But the trajectories are forming. Language dominance, training resources, network effects, and economic concentration are pushing toward outcomes that will shape civilization for centuries.

We’re not fighting over training data. We’re fighting over whose civilization’s values, assumptions, and ways of thinking become embedded in the machine intelligence that will mediate reality for our children and their children after.

That’s the data war that matters. And most of humanity doesn’t even know it’s happening.

Related Articles:

AI and Cultural Imperialism: Whose Values Get Encoded? https://www.technologyreview.com/cultural-bias-ai-training-data/

The Language Gap: Why AI Capability Determines Cultural Survival https://www.wired.com/story/language-extinction-ai-age/

The Coming Data Wars (2023 Original Column) https://futuristspeaker.com/artificial-intelligence/the-coming-data-wars/