OpenAI and xAI recently released Sora 2 AI and Grok Imagine v0.9, respectively. Both AI models are intended to create more precise and realistic videos and can replicate speech and sound effects efficiently. But with the accuracy these AI models provide, the future of entertainment may be entirely run by AI.
OpenAI, the same company behind ChatGPT, released Sora 2 on Sept. 30. The quality and sound of the videos produced by Sora 2 are far superior to those made by past models. Aside from generating realistic videos, Sora 2 is also capable of replicating animation styles and the voices of fictional characters.
However, soon after its launch, many copyright holders, such as Disney, discovered instances of copyright infringement on the app and proceeded to opt out their intellectual properties (IPs) from being utilized in prompts, according to the Forbes article “Sora 2 Does A Copyright Somersault Upon Launch,” by Dale Nelson. Still, it hints at the possibility that large corporations may be locking their IPs behind these copyright walls so that only they can generate videos of their characters. An entire future where corporations are the only ones able to profit from AI-generated videos of their IPs, and humans have no control over whatever entertainment is made, is coming into effect.
In addition, if an animation studio is the only one that can generate its IPs, they have full control over the type of content produced for said IP and can fire any unneeded artists and animators. This implies a future where companies can solely rely on AI to produce entire movies, advertisements, and cartoons. With the rapid development of AI video generation, real artists and animators may risk losing their jobs.
Corporations are even creating AI actors to replace real ones. For example, Tilly Norwood, a 100% AI actor created by a talent studio called Xicoia, is said to represent the future of the film industry, according to The Guardian article “Tilly Norwood: how scared should we be of the viral AI ‘actor,’” by Stuart Heritage.
As for Grok Imagine v0.9, audio is now synced to the generated visuals as opposed to being separated, and smoother physics were given to characters in the AI videos, according to cometapi.com. Based on my personal experience, Grok Imagine provides far fewer limitations on copyright while still being able to compete with Sora 2’s video generation. However, it fails to replicate animation and art styles as well as Sora 2, and cannot copy real-world voices like Sora 2’s “cameos” feature.
The “cameos” feature can precisely replicate a person’s voice and face and add them to an AI-generated video. The feature only requires a person to speak a series of numbers into the microphone and do a few head movements before being accessible in videos. Anyone can set their cameo to public, allowing everyone to add them to videos, but they can also limit it to friend-only use as well.
OpenAI marketed Sora 2 as a social app, so “cameos” may introduce an entirely new form of entertainment to the masses. Similar to virtual reality (VR), cameos can act as a way of immersing oneself in any situation and event, but are essentially a less interactive version of VR. Being able to add 1:1 versions of yourself and friends to absurd, realistic AI videos puts into perspective how advanced AI video generation models have become.
With Grok and Sora able to generate one’s imagination at a whim, the future of entertainment is looking to be fully AI-generated. Artists and animators may no longer be needed to draw for cartoons and animated shows, nor actors for movies, if their likenesses are already uploaded for anyone’s use on these apps. These video generation models provide infinite possibilities to users, from AI videos of oneself hanging out with celebrities to vibrant, animated fights between fictional characters.
After these models have a few more years to develop, we may have full-fledged movies being created by AI in no time.
