OpenAI announced on Thursday the launch of a new tool named Sora, which will initially have limited availability to a select group of artists, filmmakers, and “red teamers” — researchers who explore potential malicious uses of AI tools.
Sora is an advancement of OpenAI’s DALL-E, an image-generating tool. It works by interpreting prompts from users, expanding them into detailed instructions, and then utilizing an AI model trained on video and images to create new videos.
Over the past year, the quality of AI-generated content such as images, audio, and video has significantly improved. Companies like OpenAI, Google, Meta, and Stable Diffusion have been in a race to develop more capable tools and explore avenues for commercialization.
However, concerns have been raised by democracy advocates and AI researchers about the potential for these tools to be used for deception and misinformation.
