LOS ANGELES — America's entertainment industry is escalating its legal battle with artificial-intelligence companies as studios, music publishers and technology developers fight over how copyrighted material can be used to train generative models.
A new lawsuit involving major music companies and Anthropic has brought the dispute back into the national spotlight.
Who Owns the Material Used to Train AI?
AI companies need enormous quantities of information to develop sophisticated models.
That information can include books, articles, images, recordings and other creative works.
Copyright holders argue that companies should not be able to use those materials commercially without permission.
AI developers have argued that training systems on large collections of information can qualify as lawful use under copyright principles.
The issue has not been fully settled.
Recent litigation has increased the financial stakes.
Anthropic previously reached a $1.5 billion settlement involving authors who accused the company of using pirated books to train its systems.
Now major music companies are pursuing separate claims involving lyrics and sheet music.
The outcome could influence how future AI models are trained.
If courts require licenses for large portions of training data, the cost of developing AI could rise significantly.
That could benefit copyright owners while creating barriers for smaller AI companies.
If courts give developers broad protection, artists may have less control over how their work is incorporated into commercial AI systems.
Hollywood is also exploring negotiated solutions.
Technology companies and entertainment organizations are increasingly discussing licensing frameworks and safeguards.
The dispute therefore extends beyond one company.
It is becoming a fundamental question about how intellectual property works in an era when machines can absorb enormous libraries of human-created material.
For American creators, the legal decisions could determine whether AI becomes a new source of licensing revenue or a competitor built partly on their existing work.
.webp)
