Upon launch, Sora will be available to a select group of artists, filmmakers, and researchers who will explore potential malicious uses of the tool, such as spreading election misinformation and creating deepfakes. OpenAI is also developing tools to detect whether a video was generated by Sora. The company acknowledges the potential for misuse of the technology and emphasizes the importance of learning from real-world use to create safer AI systems over time.
Key takeaways:
- OpenAI has announced Sora, a new text-to-video model that can generate brief 1080p videos from typed prompts, featuring multiple characters and detailed backgrounds.
- Sora's creations are being praised for their high video-quality, consistency, style range and length, with examples including realistic and animated scenes.
- Sora can generate looping videos, edit existing videos, simulate digital worlds like video-game environments, and determine the physics of each object in a digital environment.
- Upon launch, Sora will initially be available to a small group of artists, filmmakers, and researchers focused on exploring potential malicious uses of the AI tool, such as spreading election misinformation and creating deepfakes.