Explore how major media companies like Disney and Universal are battling AI firms over copyright infringement, from training data lawsuits to billion-dollar settlements in 2025.
The artificial intelligence revolution has transformed countless industries since ChatGPT’s launch three years ago, but one area remains locked in an intense legal battle: copyright protection. As AI companies continue to train their models on vast datasets, major copyright holders are striking back with unprecedented legal challenges that could reshape the entire AI landscape.
The Copyright Battleground Heats Up
The conflict between AI innovation and intellectual property rights has reached a critical juncture in 2025. Leading AI firms have built their sophisticated models by collecting massive amounts of data from the internet, including copyrighted material, often without explicit permission from rights holders.
This year has witnessed a surge of high-profile lawsuits as major copyright owners push back against what they perceive as unauthorized use of their intellectual property.
Disney vs. AI: The Billion-Dollar Standoff
One of the most significant cases emerged in June when entertainment giants Disney and Universal filed a lawsuit against AI image generator Midjourney. The companies allege that Midjourney trained its AI models on their copyrighted content, enabling users to create images that “blatantly incorporate and copy Disney’s and Universal’s famous characters.”
The stakes couldn’t be higher, with potential damages reaching into the billions. However, Midjourney has mounted a robust defense, arguing that “the limited monopoly granted by copyright must give way to fair use” – a legal doctrine that could allow AI firms to train on copyrighted works if the outcomes are sufficiently transformative.
OpenAI’s Sora 2: Navigating Celebrity Rights
OpenAI’s latest AI video generator, Sora 2, has sparked additional controversy due to its ability to create lifelike footage of real people. The company faced criticism from representatives of Martin Luther King Jr.’s estate after the AI was used to create pastiches of his famous speeches, including inappropriate content.
In response to these concerns, OpenAI has implemented new limitations:
- Celebrities and public figures must opt out if they don’t want their images used
- Strengthened controls over depicting historical figures
- Enhanced content moderation for sensitive portrayals
The Anthropic Settlement: A $1.5 Billion Resolution
Perhaps the most significant development came when Anthropic, the company behind the Claude chatbot, settled a major lawsuit for a minimum of $1.5 billion. Three authors had alleged that Anthropic knowingly downloaded over 7 million pirated books to train its AI models.
While a judge determined that using copyrighted material for AI training might not inherently breach copyright due to its “transformative” nature, the piracy allegations were serious enough to warrant the substantial settlement.
Government Intervention and Global Perspectives
The copyright debate has attracted international attention, with governments beginning to weigh in on the issue. In October 2025, the Japanese government formally asked OpenAI to respect the intellectual property rights of Japanese culture, particularly concerning popular manga and video games published by Nintendo.
This governmental intervention signals a growing recognition that AI copyright issues extend beyond individual lawsuits to matters of national cultural protection.
Industry Expert Insights
According to Andres Guadamuz from the University of Sussex, the copyright challenges aren’t delivering the knockout punch many expected. “Many people thought that copyright would be the silver bullet that killed AI, but it is not turning out that way,” he explains.
Guadamuz predicts that AI firms will likely navigate these challenges through a combination of strategic settlements and licensing agreements rather than face existential threats to their business models.
The Fair Use Defense Strategy
The legal landscape remains complex, with AI companies increasingly relying on the “fair use” doctrine as their primary defense. This legal principle allows limited use of copyrighted material without permission when:
- The use is transformative in nature
- It serves a different purpose than the original work
- It doesn’t significantly impact the market for the original
However, the application of fair use to AI training data remains largely untested in courts, making each case a potential precedent-setter.
What’s Next for AI and Copyright?
As we move further into 2025, several key trends are emerging:
- Increased Settlements: More AI companies are choosing to settle cases out of court rather than risk potentially devastating jury verdicts
- Proactive Licensing: Forward-thinking AI firms are beginning to negotiate licensing deals with major content creators before training their models
- Regulatory Frameworks: Governments worldwide are developing specific regulations for AI training data and copyright protection
- Industry Standards: The AI industry is moving toward establishing voluntary standards for respecting intellectual property rights
Conclusion: A New Era of AI Development
While copyright lawsuits continue to challenge AI companies, experts believe these legal battles represent growing pains rather than existential threats. As Ed Newton-Rex, former AI executive and founder of campaign group Fairly Trained, notes: “AI is here to stay, even if many of the existing companies do not make it due to lawsuits.”
The resolution of these copyright disputes will likely establish the framework for responsible AI development moving forward, balancing innovation with respect for creators’ rights. As the industry matures, we can expect to see more collaborative approaches between AI developers and content creators, potentially leading to new revenue-sharing models and licensing agreements that benefit all parties involved.
The battle between AI innovation and copyright protection continues, but one thing is certain: the outcomes of these legal challenges will shape the future of artificial intelligence for years to come.
This blog post is based on research published in N New Scientist, December 2025 edition.