> ## Content Index
> Fetch the complete content index at: https://wire.fourthweb.ai/llms.txt
> Use this file to discover other available public pages before exploring further.

# USA Today Wants $250M From OpenAI for Every Article It Scraped
- URL: https://wire.fourthweb.ai/usa-today-wants-250m-from-openai-for-every-article-it-scraped/
- Published: 2026-10-08T17:58:33.000Z
- Updated: 2026-10-08T19:00:40.000Z
- Description: The training data bill is coming due, and it's being itemized line by line. USA Today Co. is suing OpenAI for over $250 million, alleging the company scraped "hundreds of thousands" of articles to train its models without permission or compensation
- Author: Travis Wright
- Tags: AI Agent Economy, AI Agents, DeFi, OpenAI, Anthropic

**The training data bill is coming due, and it's being itemized line by line.**

### The Summary

- [USA Today Co. is suing OpenAI for over $250 million](https://www.theverge.com/ai-artificial-intelligence/1008198/usa-today-openai-copyright-lawsuit?ref=wire.fourthweb.ai), alleging the company scraped "hundreds of thousands" of articles to train its models without permission or compensation
- This adds to a growing stack of publisher lawsuits including The New York Times, The Intercept, Ziff Davis, and CBC/Radio-Canada
- The legal pattern reveals a fundamental business model collision: publishers need subscriptions, AI companies need training data, and neither side has figured out how to split the value

### The Signal

[USA Today's lawsuit](https://www.theverge.com/ai-artificial-intelligence/1008198/usa-today-openai-copyright-lawsuit?ref=wire.fourthweb.ai) claims [OpenAI](https://wire.fourthweb.ai/tag/openai/) caused "real and continuing" harm to its business by training on its content without authorization. The $250 million price tag is significant, not for the dollar amount, but for what it represents: an attempt to retroactively charge for data OpenAI already ingested. The legal theory is copyright infringement. The business theory is that training an AI on journalism should cost something.

The pattern is what matters here. Every major publisher filing similar suits is creating a pricing database for training data. When The New York Times sued, the signal was uncertainty. When USA Today and a half-dozen others followed, the signal became clearer: publishers believe they can extract rent from AI companies retroactively.

> "Every lawsuit is a negotiation that started too late."

Here's the problem for both sides. OpenAI already trained its models. The content is baked in. A $250 million settlement doesn't un-train the model. It just sets a price for past behavior. Which means the real negotiation is about future access, not past scraping. USA Today isn't trying to get its articles out of GPT-4\. It's trying to get paid before GPT-5 launches.

The AI companies have three options:

- Pay publishers for ongoing access to fresh content
- Train only on licensed or public domain material going forward
- Argue fair use and take their chances in court

Most are doing all three at once. OpenAI has licensing deals with some publishers while fighting others in court. [Anthropic](https://wire.fourthweb.ai/tag/anthropic/) is exploring synthetic data generation to reduce reliance on scraped content. Google has Search, so it can offer traffic in exchange for crawl access. The model changes depending on who has leverage.

### The Implication

Watch for a wave of settlements in early 2027\. Not because anyone will win in court, but because both sides need certainty more than they need vindication. Publishers want guaranteed revenue streams. AI companies want to train without legal overhang. The settlements will set the market rate for training data, which will then determine which publishers survive and which AI companies can afford to keep scaling.

If you're building anything in the agent economy, pay attention to how this shakes out. The data licensing framework that emerges from these lawsuits will define what your agents can read, what they can summarize, and what they have to pay for. The cost of intelligence is being negotiated right now in courtrooms, not in boardrooms.

### Sources

[The Verge AI](https://www.theverge.com/ai-artificial-intelligence/1008198/usa-today-openai-copyright-lawsuit?ref=wire.fourthweb.ai)