Here comes the AI slop: OpenAI launches Sora 2 and people are already using it to create new scenes from movies and TV shows
Last year, OpenAI launched Sora, a generative AI tool that was capable of creating realistic and imaginative videos from text instructions. The results were impressive, but still looked very much like an AI creation. Now, the tech company has announced Sora 2, which is meant to be the next evolution of AI video.
Again, the results are impressive, but what’s concerning is how people are already using it to create scenes from movies and TV shows. I can’t imagine Hollywood will stand still with that. Users have been posting clips created by Sora 2 all over X, and it seems particularly adept at re-creating animation styles from shows like Family Guy, Rick & Morty, and South Park.
Other examples include individuals creating videos based on popular franchises such as Avatar, Batman, Star Wars, Super Mario, and more.
Some of the videos I’ve seen have been almost unbelievable, but they still have that AI sheen that gives them away, at least to my eye. The company has boasted that the new model is physically accurate, realistic, and more controllable than prior systems. It also features synchronized dialogue and sound effects.
“Prior video models are overoptimistic—they will morph objects and deform reality to successfully execute upon a text prompt,” reads a statement from the company. “For example, if a basketball player misses a shot, the ball may spontaneously teleport to the hoop. In Sora 2, if a basketball player misses a shot, it will rebound off the backboard. Interestingly, “mistakes” the model makes frequently appear to be mistakes of the internal agent that Sora 2 is implicitly modeling; though still imperfect, it is better about obeying the laws of physics compared to prior systems. This is an extremely important capability for any useful world simulator—you must be able to model failure, not just success.“
For better or worse, AI seems to be here to stay. It obviously poses a threat to countless jobs in the entertainment industry (not to mention many others), and some believe that the technology will allow anyone to generate a big-budget blockbuster at the click of a button. At the moment, I don’t think we’re quite there yet. While Sora 2 is undoubtedly good at generating certain elements, I’ve yet to see anything truly unique that wasn’t based on something the vast amounts of videos it’s learning from.
The post Here comes the AI slop: OpenAI launches Sora 2 and people are already using it to create new scenes from movies and TV shows appeared first on JoBlo.