As stated by Midjourney CEO and Founder, David Holz, the company will start to train datasets to create AI-generated videos. The final product will be released in a few months from now.
Midjourney enters into the video world
As reported By Decrypt, an online resource that specializes in AI and cryptocurrency, the AI image generator MidJourney will begin training its video model in the upcoming days and they expect to release a final product “in a few months”. Decrypt states: “Midjourney, the generative image creation tool perhaps best known for running inside a Discord server, is spreading its AI wings. The creators of Midjourney announced on Tuesday that they plan to introduce a ‘Text to Video’ model in the next few months. The company will begin training its video models starting in January, CEO David Holz said during an ‘Office Hour’ Discord session. This move represents a natural progression for the platform, building upon a mature image model to stir the competitive dynamics of the generative video industry”.
“Text to Video” in a few months
Decrypt mentioned that the Discord session notes included planned tweaks for V6 Niji —Midjourney’s manga/anime generator model—and consistency fixes for the upcoming official release of Midjourney V6. The company also wrote that its to-do list calls for “training for new video models to commence,” which could potentially be ready “in a few months.” Decrypt added that this venture into video also comes in the wake of releases from the competition. Stability AI recently announced Stable Video Diffusion; Meta just showcased its EMU video generator, and existing models like Pika and Runway ML are marking their territory, leaving Midjourney’s entry to emerge into a robust competitive landscape. Additionally, other image generators like Leonardo AI have already implemented video generation capabilities, further intensifying the race. Furthermore, Decrypt added the most important statement, saying that the implications of these developments extend far beyond a corporate race for supremacy. As Midjourney and others innovate and refine their offerings, the creative and media industries stand on the brink of a transformative era. The ability to generate, manipulate, and interact with video content through AI opens up many possibilities—from making things easier for entertainers and advertisers to potentially reshaping how we perceive reality.
A real threat to the filmmaking industry
As we reported previously, a class action lawsuit was filed against the main AI-imagery generator (read: Midjourney Is Being Class-Action Sued for Severe Copyright Infringements) due to the vast training that the AI-image generator executes on enormous copyrighted datasets. That was supposed to be a huge and significant lawsuit to defend artists.
The lawsuit was filed against Stability AI, Midjourney, and DeviantArt for DMCA violations, right of publicity violations, and unlawful competition. Unfortunately, in July 2023, the U.S. District Court was inclined to dismiss most of the lawsuit.
Moreover, AI-imagery generators are getting more and more popular, and implemented by industry professional tools (Adobe). But now, the situation gets even worse, as those generators enter the video world. For instance, Stable Video Diffusion is a “diffusion model for high-resolution, state-of-the-art text-to-video and image-to-video generation”. Recently, latent diffusion models trained for 2D image synthesis have been turned into generative video models by inserting temporal layers and fine-tuning them on small, high-quality video datasets. Watch the demonstration below:
Also, there’s the EMU video generator which is defined as a simple method for text-to-video generation based on diffusion models, factorizing the generation to train high-quality video generation models efficiently.
Limited solutions…for now
The ability to create videos based on trained datasets is solid, however, it’s pretty limited as well. You can explore the AI-generated videos and see it’s not there yet. You can not create a blockbuster based on AI-generated imagery (yet!). Nevertheless, the entrance of Midjourney to this party can accelerate the ability to generate very high-quality videos. And that’s a piece of bad news for our industry since AI will engineer storytelling, as stated by one of the most acclaimed Hollywood directors. Hence, filmmakers and VFX artists will need to adapt themselves and maneuver around AI to stay relevant. That would be a very challenging task.
65 seconds of the article:
