Warner Bros. Discovers Midjourney Removed Its Own Video Guardrails

The studio's new legal filing claims the AI company deliberately stripped protections that blocked copyrighted characters—then marketed it as an "improvement."

Type "classic superhero battle" into Midjourney's latest video model and watch Superman appear. Not a generic caped figure—Warner Bros. Discovery's actual Superman, complete with the iconic S-shield and red cape. According to a legal complaint filed January 8th, this isn't a bug. It's evidence of what WBD calls "systematic" and "willful" copyright infringement, a legal distinction that could cost Midjourney $150,000 per infringed work.

The complaint reveals a specific technical change that appears to have triggered this escalation: Midjourney allegedly removed video guardrails that initially blocked Warner Bros. properties when the feature launched in December. Within weeks, according to the filing, the company reframed this removal as an "improvement" that would result in "fewer blocked jobs." WBD claims Midjourney knew exactly what it was doing.

"Midjourney has made a calculated and profit-driven decision to offer zero protection for copyright owners even though Midjourney knows about the breathtaking scope of its piracy and copyright infringement," the complaint states. The language marks a sharp departure from typical corporate legal filings, using terms like "brazenly dispenses" to describe Midjourney's approach to intellectual property.

The timing matters. Disney and Universal filed their initial lawsuit against Midjourney last June. Warner Bros. followed with its complaint in September. But this January filing represents an escalation, specifically targeting the "willful" nature of the alleged infringement—a legal standard that opens the door to maximum statutory damages.

"The WBD complaint appears to more fully emphasize that Midjourney is secondarily liable for infringement via contributory, inducement and/or vicarious liability by inducing its users to directly infringe," notes Eileen McDermott, Editor-in-Chief at IPWatchdog. The distinction could prove crucial. Proving willful infringement requires showing the defendant knew about the infringement and continued anyway—exactly what WBD claims the guardrail removal demonstrates.

The stakes extend beyond Warner Bros.' character catalog. With 21 million users as of September 2024, Midjourney has become the dominant player in consumer AI image generation. Its Discord server buzzes with creators generating everything from concept art to memes. The company's move into video generation promised to expand those creative possibilities. Instead, it may have handed studios their strongest evidence yet.

"This case could set a precedent that defines how intellectual property—from fictional characters to real human faces—is treated in the age of AI," observes Biljana Cveijetinovic, a legal tech analyst quoted in the filing analysis.

Warner Bros. Discovery issued a pointed statement: "The heart of what we do is develop stories and characters to entertain our audiences, bringing to life the vision and passion of our creative partners. Midjourney is blatantly and purposefully infringing copyrighted works."

Midjourney hasn't responded to this specific filing. In previous court documents related to the Disney and Universal case, the company argued that "Copyright law does not confer absolute control over the use of copyrighted works... The limited monopoly granted by copyright must give way to fair use." Whether that defense holds when a company allegedly removes its protections remains to be seen.

The competitive landscape adds another layer. OpenAI's DALL-E 3 and upcoming Sora video model maintain stricter content filters and have pursued licensing deals with publishers like News Corp. Meanwhile, MiniMax's Hailuo AI, marketed as a "Hollywood Studio in Your Pocket," faces a lawsuit from the same studios. The message appears clear: generate our characters at your own risk.

Video creators using Midjourney should expect potential new restrictions or the return of guardrails. The $150,000 per-work penalty could create existential risk for AI companies without deep pockets. "Willful infringement" claims may become the standard legal strategy for rights holders, and licensing deals between AI companies and studios may accelerate as litigation costs mount. User-generated content platforms face increased scrutiny over their moderation decisions.

The court will likely focus on a narrow technical question: did Midjourney's removal of video guardrails constitute willful facilitation of infringement? The answer could reshape how AI companies approach content moderation, user freedom, and the fact that their most popular features often involve other people's intellectual property.