Will Artificial Intelligence Make or Break Creative Professions?

AI-created, human-edited.

On this episode of Intelligent Machines, renowned technology journalist Steven Levy outlined why authors should be compensated when artificial intelligence (AI) companies use their books for training large language models (LLMs). As AI models become more advanced by learning from high-quality sources, books offer some of the most valuable training data available. Levy addressed why a fair payment system for writers is not only feasible but crucial for the industry—and for society at large.

AI companies frequently use books to train their language models. These models generate text, answer questions, and even write stories by learning from a giant dataset drawn from the web, articles, and published works—including books. Most authors and publishers argue that this is a form of copying or “piracy” if it occurs without payment or permission. At the same time, some in Silicon Valley have claimed that making such material freely available helps AI advance for the common good.

However, as Steven Levy explained on Intelligent Machines, books are uniquely valuable: they encapsulate deep thinking, coherent language, and original ideas—exactly what high-performing models need. “Every book is a shot of a world,” Levy said, meaning that books provide a richness, nuance, and structure not found in randomly scraped web content.

The legal landscape is uncertain. Recent lawsuits (like the one against Anthropic) have challenged AI firms for using copyrighted books without compensation. While some settlements have proposed payment structures (e.g., $3,000 per book), judges have questioned whether these deals are sufficient, and what actually constitutes “fair use.”

Levy, both an author and part of the Authors Guild, explained that:

He initially saw benefits in wide AI access to knowledge, but realized scale and profit make book usage ethically different from a person reading one book at a time.

As AI systems generate real economic value, book creators should share in those profits if their work underpins the technology.

One common pushback is that “it’s too complicated” to divvy up payments to authors from model training. However, Levy pointed out parallels in the music industry, where collective licensing and royalty systems exist for streaming, radio, and public performances. These complex systems track, count, and pay creators—often through organizations like ASCAP and BMI.

Levy suggested the same could be done for books used in AI:

Transparency tools could track which books influence AI model outputs, just as streaming platforms count plays and distribute royalties.

Micro-payments or pooled fees could be allocated to authors based on how much their work contributed to the model.

Adding links or references in AI outputs could further support authors by driving sales or visibility.

Some claim the AI industry isn’t profitable enough to compensate all contributors. But Levy argued that if companies are investing hundreds of billions of dollars in infrastructure, “there’s enough to fairly pay for source material.” Importantly, if high-quality content is critical for building the best AI, companies have incentive to negotiate with creators to get the very best material.

Levy and the panel agreed that the current trajectory of AI is accelerating economic and cultural centralization in the hands of a few big tech firms. As generative AI expands, robust licensing or compensation models for content creators represent not only fairness, but also a practical way to distribute influence and economic benefit. Otherwise, the creative professions—including authors—could be squeezed out or see their value diminished.

Key Takeaways

Books are among the highest-quality data sources for training advanced AI.

Authors increasingly support compensation for use of their work in AI training.

Legal fights over “fair use” are ongoing; many want new laws or collective licensing systems.

Music industry licensing provides a workable template for how author royalties could be tracked and paid.

If AI companies want the best content, they should pay creators—and will have to, as lawsuits and public sentiment grow.

A fair compensation system could help maintain an ecosystem where high-quality writing thrives and enriches future AI.

As artificial intelligence continues to reshape content creation, ethical and economic questions around data sourcing are moving to the forefront. According to Steven Levy on Intelligent Machines, compensating authors for the use of their books is a fair, achievable goal—one that could help ensure AI’s benefits are more widely shared. Creating an industry-standard licensing structure could keep creativity, equity, and innovation alive as AI enters its next big phase.

Want more tech insights and AI deep-dives? Subscribe to Intelligent Machines for the latest interviews and analysis: https://twit.tv/shows/intelligent-machines/episodes/838