“`html
It seems like just yesterday we were all messing around with Dall-E or Midjourney, typing in bizarre prompts to see what surreal images would pop out. Suddenly, the world was awash in AI-generated art, some of it genuinely stunning, some utterly bonkers. But let’s be honest, the real holy grail for many has always been video. The idea of simply describing a scene – dragons flying over a cyberpunk city at dusk, perhaps, or a cat arguing with a toaster – and having it appear as a moving image? That’s the stuff of science fiction becoming, well, science fact. And guess who’s reportedly making a serious push into that frontier? None other than Midjourney AI.
The Video AI Storm: Midjourney Dives In
Right now, the buzz in the tech and creative worlds is palpable. We’ve seen early forays into AI video from players like Runway and Stability AI. The results have been… interesting. Often glitchy, sometimes weirdly fluid, sometimes just plain odd. Think melting faces or objects morphing in unnatural ways. But the potential is undeniable, particularly for creators looking to quickly prototype ideas or generate unique visual assets without needing vast budgets or complex animation skills.
Now, Variety has reported that Midjourney AI is throwing its hat into the ring. Apparently, they’ve got a Text-to-Video AI product in the works and it’s already in private alpha testing. For a company that built its reputation (and a massive user base) on pushing the boundaries of AI image generation, this move feels both inevitable and hugely significant. They’ve demonstrated an uncanny ability to produce aesthetically pleasing, often cinematic, images from simple text prompts. If they can translate that magic to the moving picture, we could be looking at a serious step change in what’s possible with Generative AI.
What Exactly is Midjourney Building?
Details are still a bit thin on the ground, as is typical for products in private alpha. What we know is it’s designed to be a Text-to-Video AI. You type in a description, and out comes a video. The core promise is the same one driving all generative media: democratising creation, enabling new forms of expression, and potentially speeding up production workflows significantly. Given Midjourney’s history, one might expect a focus on quality and artistic output, perhaps aiming for that distinctive, often hyper-real or painterly style that sets their image generation apart.
The big question, of course, is performance. Early AI Video tools have struggled with consistency over time, maintaining spatial relationships between objects, and handling complex motion. Will Midjourney crack this? Their image model has gone through rapid iterations, improving dramatically with each version. If they apply that same intensity to video, the results could be impressive. Imagine generating short clips, background visuals, or even entire sequences for animations or films just by describing them. The potential applications are vast, from marketing and education to entertainment and personal creativity.
Why Now? Timing is Everything
Why launch into this space now? The market is clearly heating up. We’re seeing other companies make noise, and the underlying AI technology, particularly transformer models, continues to advance. Plus, let’s not forget the elephant in the room – or rather, the multiple elephants stomping around Hollywood soundstages and boardrooms.
This move from Midjourney AI comes at a moment of intense scrutiny and legal battles surrounding Generative AI. Major players in the entertainment industry, like Disney and NBCUniversal, are involved in significant lawsuits concerning AI training data. It’s a bit like deciding to build a new, slightly taller tower in a city where half the existing skyscrapers are currently under intense legal review for their building permits. Bold? Perhaps. Strategic? Definitely.
The Shadow of the Lawsuit: Media Giants, Artists, and the AI Copyright Battle
You can’t talk about Generative AI right now without talking about the legal storm brewing, particularly around copyright. It’s a complex mess involving how these AI models are trained and what kind of data they ingest. The core argument from copyright holders, especially in the creative industries, is that their work – their movies, their TV shows, their vast libraries of images and art – is being used to train these models without permission or compensation. And frankly, who can blame them for being concerned? If an AI can generate content that looks remarkably similar to something you spent years creating, and it was trained on your work, you’d probably have some questions too.
Several significant lawsuits have made headlines regarding these issues, involving different sets of plaintiffs and defendants:
Key Copyright Lawsuits Facing Generative AI
One prominent case involves Getty Images, which sued Stability AI in January 2023, alleging the AI company illegally used millions of Getty’s stock photos to train its Stable Diffusion model without permission or license. Getty’s claim centers on the alleged direct use of their copyrighted images.
Separately, a group of artists, including Sarah Andersen, Kelly McKernan, and Karla Ortiz, filed a class-action lawsuit in January 2023 against multiple AI companies: Stability AI, Midjourney AI, and DeviantArt. Their lawsuit alleges that these companies infringed the copyrights of millions of artists by using their work, scraped from the internet, as training data for their AI image generators without consent or compensation.
Furthermore, major media companies and publishers, including giants like The New York Times, Disney, and NBCUniversal, along with groups like the Authors Guild, have filed separate lawsuits against different AI companies, primarily targeting OpenAI and Microsoft. These cases broadly allege that these AI companies used vast amounts of copyrighted text and other content from their publications, books, and media libraries to train large language models (LLMs) without permission, leading to copyright infringement. While these lawsuits focus on text-based models (often connected to image/video generation capabilities), they highlight the widespread concern across the media and publishing industries regarding AI training data.
The heart of these claims across the various lawsuits is the alleged illegal use of massive amounts of copyrighted material for training AI models. Creators and companies investing billions in content, like those with divisions such as Disney AI and NBCUniversal AI, view this as a direct threat and a violation of their intellectual property rights. The companies being sued generally argue that using publicly available data for training is transformative and falls under doctrines like fair use. Defining what constitutes “fair use” in the context of training large AI models is a central question in these AI copyright lawsuits, and their outcomes are expected to set critical precedents for the entire Generative AI industry.
Now, let’s look at where Midjourney fits into this evolving legal picture.
Where Midjourney Fits In Among the Lawsuits
Despite being a major player in the Generative AI image space and now entering AI Video, Midjourney AI has had a different status in some of these high-profile legal challenges. As noted above, they are named as a defendant in the artists’ class-action lawsuit alongside Stability AI and DeviantArt. However, they are notably not named as defendants in the Getty Images lawsuit against Stability AI or the major publisher/media lawsuits against OpenAI and Microsoft (which include Disney and NBCUniversal as plaintiffs).
Why the difference? It’s speculative, but one factor often cited is the less transparent nature of Midjourney’s operations and, crucially, its training data sources. Unlike some other AI companies that have been more open about their datasets or have faced leaks revealing their sources, Midjourney has historically been more cagey. This opacity makes it harder for specific copyright holders, like Getty or individual media companies, to definitively prove that their specific copyrighted material was used to train Midjourney’s models. While this opacity doesn’t mean their data practices are necessarily more legally compliant, it can make building a direct, traceable legal case more challenging compared to cases where datasets or scraping methods are clearer.
So, while Midjourney isn’t currently a defendant in every major AI copyright lawsuit, they are involved in significant litigation brought by artists. The legal landscape for Generative AI legal issues is still rapidly evolving, and the net could widen for any company in the space, especially as new products like AI Video launch.
Training Data, Copyright, and the Murky Waters
The question of AI training data is perhaps the most contentious issue facing the Generative AI industry. These models require colossal amounts of data to learn how to generate convincing text, images, code, and now, video. Where does all that data come from? The internet, primarily. And the internet is full of copyrighted material – photos, articles, books, films, music, artwork. It’s a global commons, but it’s also a global repository of protected intellectual property.
The AI companies argue that scraping publicly available data for training is permissible under copyright law doctrines like fair use or text and data mining exceptions that exist in some jurisdictions. Copyright holders, conversely, argue that training an AI model is a fundamentally different use case than, say, quoting a few lines from a book for criticism, and it should require licensing or permission, especially when the output can directly compete with the originals.
How Midjourney Handles Copyright
Based on public information and past statements, How Midjourney AI handles copyright in its training data isn’t entirely clear. They haven’t published details about their training datasets, which is a key point of contention for critics and copyright holders alike. This lack of transparency is part of the reason they might not be *yet* embroiled in the same high-profile lawsuit as Stability AI and Runway concerning the Getty/Disney/NBCU data.
When it comes to the *output* generated by Midjourney, their terms of service grant users certain rights, but the underlying copyright status of AI-generated work itself is also a legal grey area. Can an AI truly hold copyright? If the human prompts the AI, is the human the author? What if the AI’s output is too derivative of the training data? These are questions courts around the world are only just beginning to grapple with.
While Midjourney hasn’t been sued by Getty or the major publisher/media groups (including Disney or NBCUniversal) in *these* specific cases, their foray into AI Video using a potentially similar training methodology could well paint a target on their back in the future. As their Text-to-Video product becomes more widely used, and if its output shows clear signs of having been trained on specific, recognisable styles or content from major studios, it becomes easier for those studios to build a case. The timing of the Midjourney AI video launch right amidst this legal firestorm feels significant – they’re entering a space where the legal ground is shifting rapidly under everyone’s feet.
The Big Picture: The Future of Generative Video
Putting the legal battles aside for a moment (though they are inseparable from the technology’s future), the launch of a credible Text-to-Video AI from a company like Midjourney is a huge technical and creative milestone. Think about the possibilities: indie filmmakers creating complex visual effects on tiny budgets, animators rapid-prototyping scenes, marketers generating bespoke video content quickly, educators creating engaging visual explainers. The barrier to entry for motion graphics and simple video creation could drop dramatically.
Of course, there are challenges. Will the output be high-quality enough for professional use? Will it be controllable enough to get exactly the result you want? Early AI video models have been better at abstract concepts or short, looping clips than coherent, narrative sequences. But with Midjourney’s track record for rapid quality improvement in image generation, it’s not unreasonable to expect significant progress quickly.
And then there’s the human impact. What does this mean for the artists, animators, and VFX professionals whose jobs involve creating these very visuals? This is a genuine concern, and it’s tied directly into the legal and ethical debates around Generative AI. Will these tools be seen as aids for human creativity, or replacements for human labour? The answer likely lies somewhere in between, but the transition won’t be painless for everyone.
So, What Happens Next?
The Midjourney AI video launch is a major event. While previously expected sometime in 2025 according to earlier reports, Midjourney has now begun rolling out its initial Text-to-Video product, the V1 Video Model. The tech world will scrutinise the quality of its capabilities closely. Can it handle complex scenes? Does it produce consistent video? How does it compare to offerings from competitors like Runway or the models being developed by Google or OpenAI?
More importantly, the legal world will be watching too. Will the launch of a powerful new video generation tool from a company previously less targeted in the major copyright lawsuits change anything? Could major media companies like Disney AI or NBCUniversal AI, emboldened by their existing cases against other AI companies, turn their attention to Midjourney if its video output suggests heavy reliance on their content? The existing AI copyright lawsuit landscape is setting precedents that will affect Midjourney, regardless of whether they are currently defendants in every case.
The future of Generative AI in general, and AI Video specifically, hinges on resolving these complex legal and ethical questions around training data and copyright. The technology is clearly advancing at breakneck speed, but the legal and societal frameworks are struggling to keep up. Midjourney’s move into video is just the latest development in a story that is still very much being written, and it’s one that will shape the creative and technological landscape for years to come.
What do you make of all this? Is Midjourney’s move into video exciting or concerning? How do you think the various AI copyright lawsuit cases will ultimately play out, and what impact will they have on the future of Generative AI?
Disclaimer: This analysis is based on publicly available information and reporting as of today’s date. Legal interpretations and facts in ongoing lawsuits are subject to change.
“`
