OpenAI Sora is a new AI model that can create one-minute videos from text prompts and it is mindblowing
OpenAI, the creators of ChatGPT, have recently introduced their latest AI model that is capable of generating incredibly lifelike one-minute videos based solely on text prompts.
The new AI model, named Sora, has left everyone in awe with its remarkable capabilities. OpenAI’s objective with Sora is to train models that can comprehend and simulate real-world movements, enabling them to assist individuals in solving problems that necessitate interaction with the physical world.
Prompt: A stylish woman walks down a Tokyo street filled with warm glowing neon and animated city signage. She wears a black leather jacket, a long red dress, and black boots, and carries a black purse. She wears sunglasses and red lipstick. She walks confidently and casually. The street is damp and reflective, creating a mirror effect of the colorful lights. Many pedestrians walk about.
きた！！！OpenAIが動画生成AI「Sora」を発表。
— ChatGPT研究所 (@ctgptlb) February 15, 2024
生成動画の品質は驚愕クオリティ！圧倒的No.1。
やっぱOpenAIは凄すぎる。。
プロンプトおよび事例は以下から👇🧵
「A stylish woman walks down a Tokyo street filled with warm glowing neon and animated city signage. She wears a black leather… pic.twitter.com/fnh5dymzlU
OpenAI CEO Sam Altman showcased the power of Sora through posts on his X account. Altman invited users to provide captions for videos they would like to see, promising to create them using Sora. The outcomes he shared were astonishingly realistic, garnering a lot of attention and excitement from the platform’s users.
Furthermore, another member of the OpenAI team shared a video generated by Sora, leaving viewers astounded by its uncanny resemblance to reality.
OpenAI’s Sora model has garnered significant attention for its ability to create intricate scenes with multiple characters, precise movements, and detailed backgrounds. Not only does the model understand user prompts, but it also interprets how these elements manifest in real-world scenarios. OpenAI emphasizes that Sora possesses a deep understanding of language, enabling it to accurately interpret prompts and generate compelling characters that express vibrant emotions. Additionally, Sora can generate multiple shots within a single video, maintaining consistency in characters and visual style.
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 (@MKBHD) February 15, 2024
The newest model: https://t.co/zkDWU8Be9S
(Remember Will Smith eating spaghetti? I have so many questions) pic.twitter.com/TQ44wvNlQw
However, it is important to acknowledge the concerns raised by popular YouTuber Marques Brownlee, also known as MKBHD. He rightly points out the potential issues associated with AI-generated videos. The ease with which Sora can generate a one-minute video from simple text prompts raises the need for caution, as this technology could be easily misused. We can recall the viral AI-generated video of Will Smith eating spaghetti in early 2023, which showcased the significant leap that Sora represents. Another notable example featured Scarlett Johansson eating something. Sora’s videos exhibit a remarkable improvement in realism and believability.
It is crucial to approach the capabilities of the OpenAI Sora model with a balanced perspective, appreciating its impressive advancements while remaining mindful of the potential implications and responsible usage.
However, OpenAI is committed to discussing safety measures. Prior to incorporating Sora into OpenAI’s products, the company will implement various important safety measures. This will involve collaborating with red teamers, experts in areas such as misinformation, hateful content, and bias, who will thoroughly test the model to identify any potential vulnerabilities. OpenAI will also develop tools to identify misleading content, including a detection classifier capable of recognizing videos generated by Sora.
Furthermore, OpenAI will utilize existing safety protocols that were developed for its products using DALL·E 3, which are also applicable to Sora. For example, within an OpenAI product, their text classifier will screen and reject input prompts that violate usage policies, such as those requesting extreme violence, sexual content, hateful imagery, or celebrity likeness. The company has also established robust image classifiers to review every frame of generated videos, ensuring compliance with usage policies before granting user access.
OpenAI is actively engaging policymakers, educators, and artists worldwide to address concerns and explore the positive applications of this new technology. “We will engage policymakers, educators, and artists globally to understand their concerns and identify beneficial use cases for this new technology. Despite extensive research and testing, we cannot anticipate all the ways in which people will utilize our technology for good or misuse it. That is why we believe that learning from real-world usage is a crucial aspect of creating and releasing increasingly safe AI systems over time,” states OpenAI.
At present, Sora is exclusively accessible to red teamers for the purpose of examining crucial areas for possible issues or risks. OpenAI is additionally granting access to visual artists, designers, and filmmakers in order to gather insights on enhancing the model. The blog states, “We are sharing our research progress in its early stages to collaborate with and receive feedback from individuals outside of OpenAI, while also providing the public with a glimpse of the forthcoming AI capabilities.”
