In a series of experimental video clips created by seven artists using OpenAI's Sora generative video creation platform, a man with a balloon head appears, but that's not the most peculiar thing you'll see today.
Unlike OpenAI's ChatGPT AI chatbot and DALL-E image generation platform, the company's text-to-video tool is not yet available to the public. However, on Monday, OpenAI revealed that they have granted access to Sora to "visual artists, designers, creative directors, and filmmakers" and showcased their creations in a blog post titled "First Impressions".
Although the length of these videos varies from 20 seconds to a minute and a half, their visual effects are stunning, with most being described as abstract art. A 20-second short film by OpenAI resident artist Alex Reben explores some of his sculptures (or at least the concept of sculptures), while creative director Josephine Miller's video showcases a scene where a model blends seamlessly with seemingly transparent colored glass.
If there were an award for best entertainment, "Air Head" by multimedia production company shy kids might win. It's a short film about a man with a balloon-like yellow head, and it's quite amusing. It might remind you of an AI adaptation of the classic film "The Red Balloon," provided you can accept the premise of the boy growing up and spending the rest of his life with a red balloon.
Sora's ability to blend the peculiar balloon head with seemingly human figures and realistic environments is truly remarkable. As Walter Woodman of shy kids pointed out, "Sora is excellent at generating things that look real, but what excites us even more is its ability to create completely surreal things." Yes, it's a fun and highly surreal short film.
However, things seem to get even more peculiar.
Another video that might startle you in the middle of the night is "Beyond Our Reality" by digital artist Don Allen Stevenson III. It's like a twisted nature documentary from National Geographic, depicting never-before-seen animal hybrids such as the "Girafflamingo," a fusion of a giraffe and a flamingo, a flying pig, and an eel cat. Each one seems as if a mad scientist captured different animals, dissected them, and perfectly fused them together to create these new hybrids.
OpenAI and the artists did not disclose the prompts used to generate these videos or the effort put into transforming ideas into the final videos. Did they simply type a paragraph describing the scene, style, and level of realism and hit enter, or was it an iterative process that allowed them to seamlessly connect the man's balloon head to his shoulders or transform a bizarre rabbit armadillo into an adorable final product?
It's not surprising that OpenAI invited creatives to test run Sora. Given the powerful capabilities demonstrated by Sora, it is most likely a threat to the livelihoods of artists, filmmakers, and animators in the art, film, and animation industries. However, most people seem to believe that it is a tool that can help them develop commercial products more quickly.
Josephine Miller stated in the blog post, "This ability to quickly conceptualize at a high-quality level challenges my creative process and helps me evolve in storytelling. It allows me to translate my imagination with fewer technical limitations."
