The Data Question America Must Answer
Artificial intelligence is reshaping American life at a pace that has outrun our policy frameworks. Self-driving systems, medical diagnostics, and content recommendation algorithms now influence decisions that once required human judgment. Behind every one of these systems lies training data—vast collections of information, images, and text harvested largely from the internet, indexed by the technology companies that control the largest platforms in the world.
The question of who owns and controls that data, and who benefits from its use, has quietly become one of the most consequential governance problems facing the nation. It deserves serious conservative attention, not because Big Tech deserves punishment, but because the arrangement raises legitimate questions about property rights, corporate power, and the relationship between innovation and consent.
The Mechanics of Scale
The modern artificial intelligence boom depends on training data at an unprecedented scale. Large language models, image generators, and other systems require millions or billions of examples to function effectively. Most of that data comes from the internet—from social media posts, published articles, photographs, and other content created by millions of people. The companies training AI systems have harvested this material with minimal compensation to creators and often without explicit permission.
This is not necessarily illegal under current copyright doctrine. Courts have generally permitted the copying of large amounts of material for research and transformative purposes. But legality and fairness are not identical. A farmer can legally stand at the edge of another's field and harvest seeds blown across the property line by the wind, but we might still call it poor neighborliness.
What makes the technology company arrangement different from simple trespass is the structure of corporate power. Three or four major platforms—Google, Meta, Amazon, and a handful of others—have become the primary gatekeepers of the internet's content. They have unprecedented leverage not only over consumers but over content creators, who have few alternatives for reaching audiences. When these same companies then use the content created on their platforms to train proprietary AI systems that will compete with those creators, something important has shifted in the balance of power.
Property Rights and Consent
Conservative political thought has long emphasized the protection of property rights as a foundation for both liberty and good governance. We rightly worry when government seizes private property without just compensation. Yet we have been slower to apply the same scrutiny when corporations do something structurally similar—using others' intellectual property without compensation or meaningful consent—provided they can dress it in the language of legal routine.
This is not an argument that AI development should be halted or that all use of publicly available data should require individual permission. The transaction costs alone would make modern machine learning impossible. But it is worth asking whether our current framework adequately protects the interests of creators, and whether the concentration of this power in a handful of corporations serves the broader public interest.
A more durable arrangement would likely involve clearer rules about data use, more meaningful compensation mechanisms for creators whose work trains these systems, and a harder look at whether the current level of corporate concentration in data access serves innovation or stifles it. These are not radical propositions. They reflect a conservative concern with both property rights and the dangers of concentrated private power.
The Case for Institutional Clarity
One of the lessons of the tech regulation debate over the past decade is that neither pure laissez-faire nor heavy-handed prohibition works well. Markets work best when rules are clear, when participants know what to expect, and when power is reasonably distributed. Right now, those conditions do not clearly exist in the AI training data space.
Congress has an opportunity—and arguably an obligation—to establish clearer guardrails. This might include requirements that companies disclose how and where they source training data, frameworks for compensating creators when their work is used, or even restrictions on the accumulation of data rights in ways that foreclose competition. These are not ideologically foreign to conservatism; they reflect a concern with order, predictability, and the proper limits of corporate authority.
It is worth noting that this problem exists partly because our copyright framework was written for an earlier technological era and has not kept pace with either the scale of data use or the business models it now enables. Updating that framework is not radical—it is how law has always adapted to technical change. The alternative is to allow the current arrangement to crystallize, which risks embedding an imbalance of power that will prove difficult to correct later.
Innovation Without Consent
There is a final point worth making, one that transcends left-right divisions: innovation does not require that we ignore the interests of those whose creativity and labor made it possible. A sustainable tech ecosystem is one in which creators have both incentive and reasonable compensation for their contributions. When that breaks down—when a few companies can harvest the entire creative output of human culture and turn it into proprietary systems without meaningful benefit to those who created it—we have not achieved pure efficiency. We have created a new form of economic extraction.
The conservative response to this should not be reflexive hostility to technology or business. It should be insistence that power, whether government or corporate, operate within clear rules that respect property rights and allow for genuine consent and fair dealing. That is not anti-business. It is pro-order, pro-property, and ultimately pro-innovation, because it ensures that the system maintains the trust and participation of the people whose work makes it run.
The question of who controls the data fueling America's AI future is not settled by markets alone. It is a question of governance, and it deserves serious, principled attention from policymakers willing to think beyond short-term partisan advantage and consider what kind of innovation ecosystem actually serves the long-term health of both business and society.
