Hollywood’s AI Hypocrisy: Why Studios Are Fighting Transparency in Copyright Battles

The Fight Over Transparency and Hypocrisy

Hollywood’s biggest players, Disney, Universal, and Warner Bros., are suing AI startup Midjourney over alleged copyright infringement, but they are fiercely resisting transparency regarding their own internal AI experimentation. This resistance isn’t merely procedural; it’s the real story, hinting at a deep hypocrisy at the heart of the burgeoning AI copyright wars. Midjourney’s latest legal manoeuvre, challenging a judge’s ruling to expand discovery beyond consumer-facing AI outputs, cuts straight to this unspoken tension.

The AI image generator alleges that the studios themselves might be ‘doing exactly what they are suing Midjourney for doing’ behind closed doors. This claim reconfigures the entire legal narrative from simple infringement to a battle over industry custom and competitive practices. It suggests a desperate attempt to control the output of AI without addressing the foundational data economics that power it. This isn’t just about protecting intellectual property; it’s about controlling the technological narrative and the future of creative content generation.

While the initial lawsuits, filed ‘last year’ against Midjourney by Disney and Universal, and ‘a few months later’ by Warner Bros., centred on the generation of iconic characters like Bart Simpson and Darth Vader, the current skirmish is about data access. Midjourney argues that if studios are developing internal image-generating AI for storyboarding or ideation using unlicensed copyrighted content, it would demonstrate a widespread ‘industry custom’ that undercuts the studios’ claims of market harm.

The Data Gold Rush and IP Lockdowns

The studios, through their lead attorney David Singer, insist they ‘do not seek to stop AI technology or even shut down Midjourney’s business.’ They ‘simply want Midjourney to stop copying their movies and TV shows and to stop distributing, publicly displaying, publicly performing, and creating derivative works that include copies of [their] famous characters without authorization.’ This sounds eminently reasonable on the surface. Yet, what happens when the very mechanism for stopping such copying—litigation—reveals that everyone is dipping into the same vast well of digital culture to train models?

Midjourney’s defence of ‘fair use’ for training data, even on globally recognised IP, isn’t new in the AI space. It’s the central legal battleground that every generative AI company faces. The studios’ reluctance to disclose their full AI prompt histories and internal development workflows, which Midjourney argues would show ‘industry custom’ in using unlicensed copyrighted content for ideation, exposes the deep chasm between public posturing and private innovation within these traditional media behemoths.

This is not merely a legal tactic; it is a desperate attempt to secure the commanding heights of AI-powered content generation for themselves. By limiting discovery, the studios retain crucial opacity around their own AI strategies, potentially allowing them to leverage the very data sources they accuse others of illegally exploiting. They are effectively pushing for a two-tiered system of content licensing for AI: one for established incumbents with deep pockets, and another, more restrictive one for nimble startups.

Global Precedent and the Future of Creation

From outside the Silicon Valley echo chamber, this specific fight takes on a different hue. In Geneva or Singapore, where global IP laws often clash, the American legal system’s struggles with data licensing for AI training are viewed with a mix of fascination and frustration. While US-based reporters often focus on the immediate legal wrangling, the broader implication is how this sets precedents for content licensing globally. Are studios attempting to construct a system where they can leverage vast, unlicensed datasets for internal research and development, while smaller AI startups must navigate complex, costly licensing frameworks?

This wouldn’t be unprecedented in the history of media consolidation, but with AI, the stakes are exponentially higher, affecting everything from independent creators using open-source tools to giant media conglomerates shaping future entertainment. The request to disclose all prompts used in Midjourney, not just allegedly infringing ones, highlights the opaque nature of AI development and the powerful incentive for studios to maintain that opacity while attacking others. It suggests that the studios are not just protecting IP, but perhaps more urgently, protecting their own AI development secrets and potential future liabilities.

The studios’ argument that Midjourney’s request constitutes a ‘fishing expedition’ rings hollow when they are simultaneously accused of withholding information critical to Midjourney’s defense. This case isn’t just about copyright; it’s about control over the fundamental building blocks of future creativity and the competitive landscape of AI. The outcome will likely shape how digital rights and intellectual property are defined in an age where algorithms are increasingly involved in every stage of creative production, from initial concept to final output. This is a fight for market dominance disguised as a copyright dispute, with the integrity of open innovation hanging precariously in the balance.