In January 2026, a coalition led by Universal Music Group filed a $3.1 billion lawsuit against Anthropic, alleging the AI company downloaded over 20,000 copyrighted songs from pirate sites to train its models. A few months earlier, Anthropic had already settled a separate case for $1.5 billion over pirated books used in training data. Over 70 copyright lawsuits are now pending against AI companies in the U.S. alone.
Here is what these lawsuits are really about: someone used an input they did not pay for, and the people who produced it want the invoice.
That impulse is correct. But it is also too narrow. The missing invoice problem is much bigger than AI, and much bigger than any single country’s legal system.
The Third Input Nobody Prices
Every business prices labor and capital with obsessive precision. Wages are negotiated. Returns are forecasted. Both show up in the accounting.
But there is a third category of input in nearly every product and service on Earth. It never appears on the bill. And it does not respect borders.
Knowledge: $8.8 Trillion in Unpriced Code
A 2024 Harvard Business School study estimated the demand-side value of open-source software at $8.8 trillion. If companies had to rebuild the code they depend on from scratch, they would spend 3.5 times more than they currently do. Linux runs the majority of the world’s servers. GPS was publicly funded. The internet’s core protocols were published openly.
That code was written by people in every time zone. A developer in Nairobi, a researcher in São Paulo, a volunteer in Bangalore. Their contributions sit inside products sold by companies headquartered in San Francisco and Seattle. The value is extracted globally. The profits land in a handful of ZIP codes.
NIH funding contributed to 354 of 356 drugs approved between 2010 and 2019 — 99.4%. American taxpayers funded the research. Pharmaceutical shareholders collected the returns. People in Lagos and Lima pay full price for the same drugs.
Resources: The Atmosphere Doesn’t Stop at EU Borders
The Democratic Republic of Congo produces 74% of the world’s cobalt. That cobalt goes into every EV battery and every smartphone on the planet. The DRC raised its royalty on strategic minerals to 10% under its revised mining code. Good start. But audits between 2016 and 2018 found that mining revenues had not been disbursed to local communities as required by law. The resource is global, the royalty is national, and even the national system leaks.
The EU’s Carbon Border Adjustment Mechanism entered its compliance phase on January 1, 2026. Importers of steel, cement, and aluminum into the EU now face a carbon price. Over 12,000 companies applied for authorization. The logic is right: the atmosphere absorbs emissions, it has a limited capacity, and using that capacity should cost something.
But the atmosphere does not stop at EU borders. Eighty carbon pricing instruments now operate worldwide, generating over $100 billion in 2024 according to the World Bank. That is real money. It still covers only 28% of global emissions. The other 72% goes unpriced. And most of that $100 billion is reinvested in national budgets, not returned to the 8 billion people who share the atmosphere.
Data: $1.14 Trillion Built on Free Labor
Global digital advertising revenue hit $1.14 trillion in 2025. That industry runs on behavioral data: every search, scroll, click, and purchase teaches an algorithm what to show next. A teenager in Jakarta and a retiree in Munich generate equally useful signals. Neither is compensated.
The AI copyright lawsuits are picking at one thread. Artists and writers in the U.S. want payment for their work being used in training. But the models were trained on text and behavior from people everywhere. The extraction is global; the lawsuits are American. Even if every case is won, the settlement checks will go to plaintiffs in one jurisdiction while the inputs came from every country on Earth.
The Wrong Question
Three categories of unpriced input. Knowledge written across every continent. Resources extracted from specific geographies but consumed globally. Data generated by billions of people in every country. All of them show up in the final product. None of them show up on the invoice.
This is where UBI funding conversations get stuck. Supporters say: give people cash. Opponents ask: who pays? The debate cycles through income tax, wealth tax, deficit spending, and budget cuts. Each depends on one government’s political mood, and each shrinks when that country’s economy contracts.
But the question itself might be wrong. Instead of asking “how do we redistribute what has already been earned,” ask: what was never priced in the first place — and who does it actually belong to?
Fragmented but Already Working
The fix is already happening. It is just fragmented and national.
Switzerland charges CHF 120 per tonne of CO₂ on heating fuels and redistributes two-thirds equally to every resident. Alaska collects petroleum royalties and pays every resident a dividend: $1,000 per person in 2025. California runs cap-and-trade auctions and sends households a Climate Credit.
Each one proves the principle works: price the shared input, return the proceeds to people. But each one stops at its own border. The Swiss dividend is for Swiss residents, the Alaska check is for Alaskans, and the California credit is for Californians.
The atmosphere does not belong to Switzerland. Cobalt does not belong to whichever company gets the mining permit. Open-source code does not belong to the country where the server is located. These inputs are universal. The returns should be too.
That does not mean it has to happen all at once. Every national system that prices a shared input and returns the revenue to people is a proof of concept. The question is whether we keep treating these as isolated national experiments — or recognize them as pieces of the same invoice.
