Apple Has a New Open-Source AI Image Editor

Apple has released MGIE, an open-source multimodal AI model for image editing, signaling a strategic push to catch up in the generative AI race ahead of iOS 18. By translating natural language instructions into specific visual edits, MGIE allows users to perform complex adjustments—from removing objects to changing backgrounds—using simple, conversational commands. This development highlights Apple’s intent to integrate accessible, intuitive AI tools into its ecosystem, potentially redefining how users interact with photo software. The technology relies on a large language model capable of interpreting both text and visuals, solving the traditional challenge of understanding vague user intent. It supports three distinct editing categories: precise modifications like cropping, global optimizations such as brightness adjustment, and localized changes to specific areas like shape or color. This versatility demonstrates a significant leap in AI’s ability to comprehend and execute nuanced creative requests, bridging the gap between casual user expectations and professional-grade editing capabilities. This release is highly relevant to open data as it provides a transparent, accessible framework for understanding how multimodal AI processes visual data. By making the model open-source, Apple contributes to the broader community’s ability to study, audit, and build upon these technologies. This openness encourages innovation and ensures that the methodologies behind AI-driven image manipulation remain visible and debatable, fostering a more transparent landscape for visual AI development and integration.

Source: lifehacker.com
Published on 2024-02-08

Anuncia INEGI Censo Económico de 2024 »

Open Data

Related news