Apple’s latest prototype AI tool can animate images using text descriptions

Apple’s latest prototype AI tool can animate images using text descriptions

Keyframer, a novel animation tool from Apple, relies on large language models (LLMs) and is designed to democratize animation by allowing users to articulate animations through text descriptions rather than complex coding. Leveraging OpenAI's GPT-4 as its backbone, Keyframer simplifies the animation process, making it accessible to users without coding expertise. This prototype demonstrates how generative AI can be applied to animation, offering a user-friendly approach to creating animated content.

The tool's functionality is centered around Scalable Vector Graphic (SVG) files, ensuring that illustrations can be resized without compromising quality. By using a text prompt, users can instruct Keyframer to animate images in specific ways, such as making stars twinkle or altering background colors. Notably, Keyframer converts these text prompts into CSS code automatically, making it easy for users to make adjustments without delving into intricate coding structures.

While the current version of Keyframer has limitations, such as a focus on web-based animations and a reliance on simple SVG images, it represents a step toward simplifying animation creation for a broader audience. The tool's potential impact on employment within the animation industry is acknowledged, with professionals viewing it as a valuable addition to their toolkit rather than a direct threat to their roles.

Keyframer joins other recent AI innovations from Apple, including Human Gaussian Splats (HUGS), capable of generating animation-ready human avatars, and MGIE, an AI model for image editing using text-based descriptions. Although Keyframer is still in its prototype stage and not publicly available, it underscores Apple's commitment to exploring the possibilities of generative AI in various creative domains.