Stability AI launches AI-powered video creator

The developer says interested users can wait for access to an 'upcoming web experience featuring a text-to-video interface'
A representational image, generated by Stable Video Diffusion, illustrates two birds. — Stabilityai
A representational image, generated by Stable Video Diffusion, illustrates two birds. — Stabilityai

Stability AI, the developer behind Stable Diffusion has announced Stable Video Diffusion, a pioneering AI-driven tool capable of generating high-quality videos using still images.

The step taken by AI startup marks a significant advancement in the realm of AI-powered video generation which has been the central stage for some of the biggest players in the field of artificial intelligence, TechCrunch reported.

This innovative tool allows users to transform a single image into a video, ranking among the handful of those that achieved this feat. However, currently, the tool is only available for research purposes.

"This state-of-the-art generative AI video model represents a significant step in our journey toward creating models for everyone of every type," Stability AI stated.

Glenn Gabe, an SEO Consultant at G-Squared Interactive focused on Google algorithm update recovery, posted on X (formerly Twitter) about the release of Stable Video Diffusion, expressing excitement and appreciation through his tweet.

The recently launched tool has two models that are capable of generating videos from 14 to 25 frames, with a speed between 3 and 30 frames per second at a resolution of 576 x 1024.

Comparing its innovative tool with text-to-video platforms Runway and Pika Labs, the company said, "At the time of release in their foundational form, through external evaluation, we have found these models surpass the leading closed models in user preference studies."

According to Stability AI, interested users can wait for access to an "upcoming web experience featuring a text-to-video interface."

The tool will showcase its potential applications across various sectors, including advertising, education, and entertainment.

The tool, generating relatively short videos (less than 4 seconds), lacks perfect photorealism has reportedly failed to control text, and cannot accurately generate people and faces.

The tool has been trained on a dataset based on millions of videos, with fine-tuning on a smaller set. The company emphasised using publicly available video data for research purposes.

The company was recently sued by Getty Images for scraping its image archives.

Unlike OpenAI's success with ChatGPT, Stability has faced challenges in commercialising Stable Diffusion, leading to massive burnout of investment.