DealMakerz

Complete British News World

The short film about the rainforest was produced entirely using artificial intelligence – are you being scammed?

The short film about the rainforest was produced entirely using artificial intelligence – are you being scammed?

The 8-minute video was created using the Stable Diffusion AI model and shows birds, water, plants and vegetation in high detail.

Creating images using AI and so-called “instant photography” is not a problem for most people – for example via Adobe Firefly, Midjourney or Stable Diffusion.

However, creating high-quality and detailed video materials with the help of AI is quite tedious. The models that generate the images consume processing power, and in addition, it takes time to render the sequences to see the result and then return them with new changes and improvements.

Still from the movie

Still from the movie

Still from the movie.

Still from the movie.

User 909Upload spent time creating an 8-minute video in Stable Diffusion:

“This is a project I started to test SVD weights in Stable Diffusion (Comfy). Visually, everything is rendered locally on my machine (RTX 3090Ti) using a variety of different tools and software for image creation, image animation, frame interpolation, and post upscaling (I found that Better than running from the lower level to the higher level of rent reform – a lot of detail is lost).”

The model rendered sequences at 1536 x 640 at 8 fps, which were then interpolated up to 3840 x 1620 at 48 fps. Speaker audio is generated using Chat GPT.

Also check out the clip below from ActionMovieDad on

“This literally started with me seeing if the new Midjourney could handle hands, and then putting hands in the water, and it spiraled out of control. I promise this is as hard as I'll ever lean into this classic men's look — but I'm also like how… it turns out.”

The graphics fool you

Game engines, the basic software that handles displaying and creating graphics, motion, and objects in a computer or video game, have become so powerful today that with today's hardware they can fool anyone into thinking it's a photograph and not a screenshot from a game. – In the same way that generative AI can fool you with its sequences.

For example, take a look at the Matrix Awakens demo and compare it to this rainforest sequence.