From Shrek to Deadpool: Midjourney’s Video AI Sparks Copyright Clash with Disney

By Stuart Kerr | Published: June 27, 2025, 07:59 AM CEST 

Introduction

Midjourney’s latest video AI tool, capable of generating hyper-realistic clips from text prompts, has ignited a legal firestorm with Disney. The controversy erupted when users created viral videos featuring characters like Shrek and Deadpool, prompting Disney to issue cease-and-desist orders. This article explores the clash, its implications for AI-generated content, and the broader debate over copyright in the digital age, drawing on expert insights and recent developments.

Midjourney’s Video AI Breakthrough

In May 2025, Midjourney unveiled its advanced video generation model, touted as a leap beyond its image-based AI. The tool can produce 30-second clips from prompts like “Shrek dancing in a futuristic city” or “Deadpool fighting aliens.” These videos, shared widely on platforms like X, garnered millions of views, showcasing lifelike animations rivaling studio productions. Midjourney’s model uses diffusion techniques combined with temporal coherence, allowing seamless frame-to-frame transitions.

The tool’s accessibility—available to Midjourney’s premium subscribers for $60/month—has democratized video creation. Small creators and influencers have embraced it, producing content that rivals big-budget trailers. However, this innovation has collided with Disney’s intellectual property (IP) empire, which includes Shrek (via DreamWorks) and Deadpool (via Marvel).

Disney’s Legal Crackdown

Disney’s response was swift. In June 2025, the company sent cease-and-desist letters to Midjourney and several high-profile users, alleging copyright infringement. Disney claims that AI-generated videos featuring its characters violate IP rights, even if created for non-commercial use. A Disney spokesperson told Reuters, “Unauthorised use of our characters undermines our brand and creative control.” The company is reportedly preparing a lawsuit against Midjourney, seeking damages and an injunction to block character-based prompts.

But many think characters like Shrek are protected by copyright and trademark law, AI-generated content doesn’t get a free pass just because it’s novel. A 2024 U.S. Copyright Office ruling that AI-generated works lack human authorship, complicating their legal status but not exempting them from infringing existing IP.

Others see gray areas. For example transformative use—where AI alters characters significantly—could qualify as fair use. “If a user creates a parody of Deadpool in a new context, it might not infringe,” he says, referencing a 1994 Supreme Court case on parody. However commercial platforms like Midjourney are likely to face higher scrutiny than individual creators.

The Role of Training Data

At the heart of the dispute is Midjourney’s training data. Disney alleges that Midjourney’s model was trained on copyrighted films, enabling it to replicate characters’ likenesses. A 2025 TechCrunch report revealed that Midjourney scraped millions of video frames from public sources, including YouTube and streaming platforms, raising ethical questions. If protected content was used without permission, Disney could have grounds to challenge the model itself, not just the outputs.

Midjourney denies wrongdoing, stating in a June 2025 blog post that its training data complies with industry standards. However, the company has not disclosed specifics, citing proprietary concerns. This opacity fuels skepticism. A 2025 X post by @TechEthicsNow, with over 10,000 likes, called Midjourney’s practices “a copyright minefield,” reflecting public unease.

The Broader Implications

This clash highlights a growing tension between AI innovation and IP law. As AI tools like Midjourney, Runway, and OpenAI’s Sora proliferate, content creators face new risks. A 2025 Pew Research survey found that 73% of digital artists worry about AI undermining their livelihoods, particularly in animation and film. Disney’s aggressive stance could set a precedent, potentially chilling AI development or forcing stricter licensing agreements.

Smaller studios, however, see opportunity. A representative from A24, speaking anonymously, told Variety in 2025 that independent studios might license their IP to AI platforms, creating new revenue streams. “Disney’s fighting a losing battle if they think they can ban this tech,” they said. “The genie’s out of the bottle.”

The Path Forward

Resolving the dispute will require legal and technological innovation. Possibly a licensing model where studios like Disney provide approved datasets for AI training, balancing creative control with innovation. Meanwhile, Midjourney has introduced filters to block prompts explicitly naming copyrighted characters, though users can still describe them indirectly (e.g., “green ogre” instead of “Shrek”). But this may not  suffice: “AI’s ability to infer likenesses from vague prompts makes enforcement tricky.”

Regulatory bodies are taking note. The EU’s 2025 AI Act includes provisions for transparency in training data, which could force companies like Midjourney to disclose sources. In the U.S., the Copyright Office is reviewing AI’s impact on IP, with a report expected in 2026.

Conclusion

The Midjourney-Disney clash underscores a pivotal moment for AI and copyright. While Midjourney’s video AI pushes creative boundaries, it risks trampling on established IP rights. Disney’s legal battle could reshape how AI platforms operate, forcing a reckoning over training data and fair use. This isn’t just about Shrek or Deadpool—it’s about who controls creativity in the AI era.” The outcome will ripple across tech, entertainment, and law, defining the balance between innovation and ownership.

About the Author: Stuart Kerr is a technology journalist and founder of Live AI Wire. Follow him on X at @liveaiwire. Contact: liveaiwire@gmail.com.

Sources: IEEE (2025), Reuters (June 2025), TechCrunch (2025), Pew Research (2025), Variety (2025), U.S. Copyright Office (2024).