OpenAI, the same company that made ChatGPT, launched Sora 2 on October 1st. Sora is their video creation AI model, which was first launched in February 2024. Sora 2 is very different from the original model. If the original Sora was just an AI video generator,
Sora 2 is more of a TikTok of AI videos. Sora 2’s videos are pretty realistic until something bizarre happens or you notice that the characters look a little animated. Currently, there’s a waitlist to even get on the platform, with only those given an access code allowed to use it. Also, as of the most current update, Free users can make 15-second videos, while Pro users can make up to 25-second videos. Users have already flooded the platform with their creativity, from videos of a cat being told they are not the father to those with levitating people. Despite how amusing these videos may be, Sora 2 may face some legal challenges.
Initially, Sora 2 didn’t restrict the content on the platform, so in the first 48 hours, the app was flooded with videos of Mario escaping from the cops, Pikachu committing crimes, etc. All of which infringed copyright law. So then, on October 3rd, OpenAI gave
companies the option to opt out of the model using their brand. Even then, users have been able to bypass the restrictions with sneaky prompts. A little sickening, however, is that people have been able to make videos of other things not restricted by copyright laws, long since passed historical figures, for example. The families of deceased celebrities are now getting AI videos of them made using Sora 2. Martin Luther King Jr can be seen reciting a modified version of the famous “I have a dream” speech. OpenAI had to manually prevent users from making such videos.
Now, what does the launch of Sora 2 mean for the film and content-creating industry? Sora can create incredible visuals, like those used in transition scenes in movies and YouTube videos. So, VFX artists, YouTube video editors, and other such professions could be put out of business. There are also implications that these kinds of models could replace actors. Have you heard of Tilly Norwood? If not, she’s an AI actress made by Particle6, a British company. She hasn’t starred in any major films yet, but she has done many short films. Her very existence and her realness scare Hollywood. Emily Blunt, for example, has said, “No, are you serious? That’s an AI? Good Lord, we’re screwed.” AI actors are sure to make the cost of creating a film a lot cheaper. OpenAI clearly has intentions to enter the industry after launching the Storyboard feature recently. Storyboard is a feature that allows Pro users to plan out their video before making it, similar to what filmmakers do.
Of course, Sora 2 isn’t the only AI video-generating model. Google just launched Veo 3, their own AI video creator. As for which model is better, it is subjective. While OpenAI’s model is more focused on being an AI TikTok, Google’s model is more focused on longer videos and creating tools to help filmmakers. Flow, another AI model launched by Google alongside Veo 3, is meant to be a tool to make filmmaking a lot easier.
AI models can now make videos and recreate moments however one person wants to imagine it, so just imagine what AI could do next. AI can’t control its own fate, what it knows, or what it learns from. What it creates is in our hands, so choose wisely.