After the AI Copyright Wars

While lawyers fight over old data, AI’s next battlefield isn’t books or code—it’s your daily life.

Generative AI was fueled by a gold rush for high-quality data. But those streams are nearly dry, and while lawyers fight over scraps, the industry is already moving on to a far bigger prize: your lived experience. This is because the largest language models have strip-mined the public web, books, news archives, and open-source code. Experts estimate that relatively soon, most open, high-quality material will likely be depleted, forcing a sea change in discovery, licensing, or synthesis just to keep training the models.

This data grab hasn’t gone unchallenged. In 2025 alone, over fifty AI-related copyright lawsuits were active. Judges in different circuits are split: some have found aspects of training to qualify as fair use, especially when the use is transformative; others side with creators when substitution is clear or economic harm is demonstrated. The biggest cases against OpenAI, Meta, and Anthropic continue to grind forward, with industry and plaintiffs alike waiting for a decisive precedent.

But while many eyeballs are focused on these legal outcomes, the AI industry’s data and model-training focus is already shifting to new frontiers. Why? Because the well is nearly dry and pipelines to scrape the remaining data and ingest real-time streams are already in place. These copyright wars may set legal precedent, but function mostly as rearguard actions. The true frontier has shifted from static corpora to lived experience.

Meta’s new Ray-Ban glasses push AI into everyday perception and memory. Apple’s Vision Pro bets on spatial computing as a workflow foundation. Perplexity, Grok, and others pitch conversational agents as always present, always adaptive “companions”. As the Financial Times recently put it, “experience is now the dominant medium of technological improvement”.

Critics call some of this hype, and they’re not completely wrong, but the market direction is unmistakable. The most valuable data won’t be yesterday’s text or images. It will be the messy, evolving context of ordinary life: what individuals see, hear, say, and do, captured, synthesized, and fed back in real time. Copyright wars will end in the courts, but the experience wars will be fought on our faces, in our ears, and in the human loop of daily action. The decisive battles for AI’s future won’t be over library books—they’ll be over reality itself.

This shift raises questions we're barely beginning to ask. Unlike the copyright battles fought between corporations and creators over past works, the experience wars will be waged with ordinary people as both the battlefield and the prize. When AI systems learn from how we navigate our kitchens, react to conversations, or pause before making decisions, traditional notions of consent become murky.

We're rapidly moving from a world where AI companies ask forgiveness for scraping public data to one where they'll need permission to scrape our lives. If that notion is correct, today’s book and newspaper copyright cases will look quaint compared to the governance challenges of tomorrow’s ambient intelligence, when the prize isn’t past content but our lived reality.