Google News
Meta, the social media behemoth, has announced the development of two groundbreaking artificial intelligence (AI) models dedicated to content editing and generation, marking a significant step forward in digital creativity tools.
Innovative AI Tools for Video and Image Editing
The first of the two models, named Emu Video, builds upon Meta's existing Emu model and is capable of generating video clips from text and image inputs. This model represents a leap in video content creation, offering creators, artists, and animators a powerful new tool for bringing their visions to life.
Meta's approach to training Emu Video involves a “factorized” method. This process is divided into two stages: initially generating images based on a text prompt, followed by creating videos conditioned on both the text and the generated image.
This method enables efficient training of video generation models, allowing for the production of 512x512 four-second-long videos at 16 frames per second. The second model, Emu Edit, focuses on image manipulation, promising enhanced precision in editing tasks.
Users can expect capabilities such as background removal or addition, color and geometry transformations, and both local and global editing of images. Meta emphasizes the model's ability to precisely alter only the pixels relevant to the user's editing request, ensuring that the integrity of the original image is maintained.
Training and Ethical Considerations
Meta trained Emu Edit using a dataset of 10 million synthesized images, each accompanied by a task description and a targeted output image, making it one of the largest datasets of its kind.
The Emu model itself was trained using a staggering 1.1 billion pieces of data, including photos and captions shared by users on Facebook and Instagram, as revealed by CEO Mark Zuckerberg. Despite the potential of these AI tools, Meta is proceeding with caution due to close regulatory scrutiny.
The company recently announced that it would not allow political campaigns and advertisers to use its AI tools for ad creation on Facebook and Instagram. This decision reflects the broader context of ethical considerations and regulatory compliance in the deployment of AI technologies.
