The news: Disney and Universal are suing AI startup Midjourney, alleging copyright infringement.

The lawsuit states that Midjourney pirated their content libraries and continues to produce “innumerable” copies of their characters, including Shrek, Homer Simpson, and Darth Vader.

The case marks the first time Hollywood giants have gone after an AI company, per NPR, following lawsuits like The New York Times suing OpenAI and Microsoft and major music companies suing Suno and Udio.

The background: Midjourney generates images from typed user queries and serves as a small competitor to tools like OpenAI’s DALL-E. Over half (53%) of marketers worldwide use visual AI tools like Midjourney to create social media content, per HubSpot.

Protecting creative labor: This lawsuit is another step in the creative industry’s efforts to protect intellectual property from AI companies.

“We are bullish on the promise of AI technology … but piracy is piracy, and the fact that it’s done by an AI company does not make it any less infringing,” Disney chief legal and compliance officer Horacio Gutierrez said, per CNN.

Why does it matter? The case could raise the stakes for any business that’s developing or deploying generative AI (genAI) tools and set precedents around training data and fair use.

It also highlights a major challenge for AI companies: The shrinking pool of publicly available data for model training.

• If Disney and Universal win their case, restrictions around IP use for AI development could make it more difficult—and more expensive—to train new models without licensing deals.
• It could also give media companies more leverage in such deals and raise their value if AI companies have no choice but to pay for content.

Our take: A victory for the studios could cut off AI companies’ access to media libraries, accelerate a shift toward paid content licensing deals, and set legal precedents to help web publishers and IP owners protect their content from data scraping.

Marketers that are reliant on genAI tools like Midjourney could see model development slow if training data resources dry up and media companies aren’t willing to share content.