Runway Releases AI Model Capable Of Videos From Text Description

Runway, a startup that helped create the popular AI image generator Stable Diffusion. (photo: Twitter @Contary_Res)

JAKARTA - Runway, a startup that helped create the popular Artificial Intelligence image generator Stable Diffusion, has released an AI model that can take any text description, such as "turbes flying in the sky", and produces three seconds of appropriate video recordings.

Citing security and business reasons, Runway has not released the model widely for now, nor will it be open-source like Stable Diffusion. The text-to-video model called Gen-2 will initially be available on Discord via a waiting list on the Runway website.

Using AI to produce videos from text input is nothing new. Meta Platforms Inc and Google both also released research papers on text-to-video AI models late last year. "The difference is that Runway's text-to-video AI models are made available to the general public," said Cristobal Valenzuela, Runway's chief executive.

"Runway hopes creatives and filmmakers will use this product," said Valenzuela.

Stable Diffusion is an AI image generator that uses a generating algorithm to generate realistic and high-quality images. Runway is likely to use machine learning technology to train the AI Stable Diffusion model and develop algorithms used in those models.

Once the model is running, users can enter text or image inputs and produce new images based on these inputs.

Several companies other than Runway also have similar technologies for producing videos or images from text descriptions through the use of AI are:

OpenAI: This leading AI research company has released several AI models capable of producing images and text from certain descriptions, such as DALL-E and CLIP.

Google: Google also has similar technologies, such as AI Back Screen launches that can generate videos from text descriptions.

NVIDIA: This technology company is developing AI technology to create realistic 3D models from 2D photos.

IBM: IBM is also developing AI technology to produce realistic images known as "Generative Adversarial Networks" (GANs).

However, it should be noted that AI technology for producing videos or images from text descriptions is still in the development stage and some of the models released by these companies may not be available to public users.