- By quade
- 20 February 2024
Sora AI Model Can Create Short Form Videos Through Text Prompts.
OpenAI, the creator of products such as ChatGPT and DALL·E, has developed its latest AI product, Sora. This product aims to create short-form videos of about a minute long with just a single text prompt. AI content has spread rapidly throughout the last few years, but so far AI videos have been the most difficult to create. While there already have been AI-generated videos, many of these appear awkward and unrealistic.
Sora hopes to change that and create content that is high-quality, realistic, and entertaining.
A New Level of Realism?
All videos by Sora will be short, only about a minute long but they can be filled with content. OpenAI claims that Sora can create scenes featuring multiple characters, carefully created backgrounds, and realistic movements.
This is because the AI is programmed to not just understand the user’s prompts, but also the real-world physics that come into play during these scenarios. Sora will not just create new videos, their goal is to also make these videos as realistic as possible. According to the Sora blog, they have trained the model to not just follow the prompts, but understand the physics in place. This will create more realistic situations, backgrounds, and effects during these videos.
With this extra training, OpenAI hopes to create videos that can fully mimic the effects of the real world.
“We’re teaching AI to understand and simulate the physical world in motion, with the goal of training models that help people solve problems that require real-world interaction,”
-The OpenAI Sora blog.
OpenAI CEO Sam Altman also showed off the tool via posts on his X account. “We’d like to show you what Sora can do, please reply with captions for videos you’d like to see, and we’ll start making some!” Altman wrote in his post. A lot of users on the platform sent him prompts, and the results he shared look very realistic.
Among the most excited about this reveal was Sam Altman, the CEO of OpenAI. In a post talking about Sora, he invited people to comment on text prompts. The best prompts would be made into videos that would be posted.
“The model has a deep understanding of language, enabling it to accurately interpret prompts and generate compelling characters that express vibrant emotions. Sora can also create multiple shots within a single generated video that accurately persist characters and visual style,”
Statement by OpenAI.
Safety Concerns About Sora
While many are excited about Sora, others have expressed concern about the dangers of this tool. Users noted the infamous Will Smith eating spaghetti video. While the improvements were impressive, they also posed a new danger. As AI becomes more advanced, the risks of people misusing it also increase as it can be used to create videos about anyone without their consent. Things like porn, propaganda, and scams are among the few risks highlighted by critics of this technology.
“Every single one of these videos is AI-generated, and if this doesn’t concern you at least a little bit, nothing will,”
-Marques Brownlee aka MKBHD, YouTuber.
However, OpenAI has gone out of its way to assure users of the safety of their product. The company says that along with improving its realism, they are also testing and implementing guard rails that will mitigate the harmful effects.
Many of these guardrails have been carried over from OpenAI’s other flagship product, DALL·E as they share similar vulnerabilities. One example is a guardrail that prevents users from inputting anything involving violence, sexual content, hateful imagery, or images using celebrity likenesses. These guardrails have allowed OpenAI to develop robust image classifiers that ensure the policy is used correctly.
They have partnered with misinformation experts, educators, artists, and other concerned parties to identify and solve potential vulnerabilities. These will help prevent users from exploiting this tool.
“We’ll be engaging policymakers, educators, and artists around the world to understand their concerns and to identify positive use cases for this new technology. Despite extensive research and testing, we cannot predict all of the beneficial ways people will use our technology, nor all the ways people will abuse it. That’s why we believe that learning from real-world use is a critical component of creating and releasing increasingly safe AI systems over time,”
– OpenAI Statement
For that reason, Sora will only be available for red teamers who are responsible for checking these potential issues. Once they address these issues, OpenAi will then open access to artists and filmmakers who can provide their input.
Our team at geniusOS believes that this application could be a game changer not only in video production but also in our services. With the opportunity to create videos through prompts, we can provide more services when developing applications.
