Google Photos to Introduce Transparency with AI-Edited Image Labeling

With high increase in use of AI-edited photos, distinguishing between original and AI generated image is becoming increasingly difficult, often requiring detailed examination. Through a recent blog, Google Photos announced a new feature to help tackle this issue. Which confirms the addition of an “AI Info” label within the metadata of images. This label will indicate if an image was generated or edited using artificial intelligence (AI), specifically through Google Photos tools like Magic Editor and Magic Eraser.

Adopting International Standards

This new feature will be in sync with standards from The International Press Telecommunications Council (IPTC), signaling generative AI usage in photos. When users view a photo in the Google Photos app and scroll to the bottom of the ‘Details’ section, a notification will appear if the image was “Edited with Google AI.”

Representational image (Cup of Couple – Pexels.com)

Detailed Metadata and Digital Source Type Labels

The metadata will provide credits to the specific tool used to modify the image, such as Magic Editor or Zoom Enhance. Additionally, a ‘Digital Source Type’ label will reveal if generative AI or other editing tools were applied. Alterations made with other tools, like Pixel 8’s Best Take and Pixel 9’s Add Me, will also include detailed information within the “AI Info” label for easy identification.

Commitment to Responsible AI Usage

Google emphasized the importance of transparency and adherence to its AI Principles as it rolls out these tools more widely. Starting next week, Google Photos will display a label whenever a photo is edited with AI-powered tools. “As we bring these tools to more people, we recognize the importance of doing so responsibly,” stated John Fisher, Engineering Director for Google Photos.

Industry-Wide Efforts for AI Transparency

This feature from Google aligns with efforts across the tech industry to increase transparency around AI-edited content. OpenAI, Adobe, Microsoft, Apple, and Meta are also exploring methods for identifying AI-generated images. For instance, Meta is transitioning from general “Made with AI” tags to “AI Info” labels that specify the AI tools used, while Apple’s upcoming Apple Intelligence features are expected to add metadata to AI-edited images to simplify identification.

Google’s ongoing commitment includes gathering user feedback and considering additional measures to enhance transparency around AI image editing.

What is Metadata? – Learn Here from Wikipedia