ChatGPT Founder OpenAI ready to follow AI Video generators Like Runway, Meta and Google.
Today OpenAI just unveiled its new text based AI Video Model “SORA” that can generate HD Quality videos with just text commands. OpenAI claiming With given details, description or still image SORA can generate 1080p movie-like scenes with multiple characters, different types of motion, complex scenes, and background details.
Sora also capable of “Extend” Video clips, which is best to fill the missing details.
OpenAI writes in a blog post that “Sora has a deep understanding of language, enabling it to accurately interpret prompts and generate compelling characters that express vibrant emotions. The model understands not only what the user has asked for in the prompt, but also how those things exist in the physical world.”
So, on OpenAI’s demo page for “Sora”, they’re using big words to talk about how great it is—like the statement above. But when you check out the examples they’ve chosen from the model, it does seems pretty cool. It looks too better than other AI text-to-video generators we’ve come across, at least from what we’ve seen.
At this time, Sora has the ability to generate videos in various styles, such as photorealistic, animated, or black and white, with a duration of up to one minute. This is a lot longer than what you typically see in other text-to-video AI models. What’s impressive is that these videos generally make sense—they don’t always have those strange moments where objects move in impossible ways, which tends to happen with some other AI models.
Here is some gif samples created from Sora generated Videos (Video Clips Provided by OpenAI)
OpenAI also stated that “The current model has weaknesses. It may struggle with accurately simulating the physics of a complex scene, and may not understand specific instances of cause and effect. For example, a person might take a bite out of a cookie, but afterward, the cookie may not have a bite mark.
The model may also confuse spatial details of a prompt, for example, mixing up left and right, and may struggle with precise descriptions of events that take place over time, like following a specific camera trajectory.”
OpenAI also claimed in its blogpost that Ahead of making Sora available in OpenAI’s products, several important safety steps will be taken. Red teamers-domain experts in areas like misinformation, hateful content will be enlisted to adversarial test the model.
Safety and Concerns about Sora AI
OpenAI says that Tools are being developed to aid in the detection of misleading content, including a detection classifier capable of discerning videos generated by Sora. OpenAi has also plan to include C2PA metadata in this model before model available in OpenAI products.
For instance, within an OpenAI product, a text classifier will scrutinize and reject text input prompts violating usage policies, such as those requesting extreme violence, sexual content, hateful imagery, celebrity likeness, or the intellectual property of others. Robust image classifiers have also been developed to scrutinize every generated video frame, ensuring adherence to usage policies before user presentation.
OpenAi quote in its blog regarding safety and concerns that “Engagement with policymakers, educators, and artists worldwide is planned to understand concerns and identify positive use cases for this new technology. Despite extensive research and testing, not all beneficial or abusive uses can be predicted. Hence, learning from real-world application is deemed a critical aspect of creating and releasing increasingly safe AI systems over time.”
The release date for “Sora” has not been confirmed yet. Whenever we receive an update, we will update you on this website or on our social media handles.
What do you think about this new Generative AI model let me know in comments below.
Follow Cybersundar for more updates:
Facebook: http://facebook.com/cybersundar
Twitter: http://twitter.com/cybersundar
Instagram: http://instagram.com/cybersundar
Want to secure your Smart Home? Learn here How you can protect it.
