Meta is set to change the label on images generated or edited using artificial intelligence (AI) from ‘Made with AI’ to ‘AI Information’. This shift aims to provide more detailed information about the technology’s usage and whether it involved minor modifications.
In February, Mark Zuckerberg-led Meta announced its commitment to identifying realistic content created by AI tools on its social networks – Facebook, Instagram, and Threads. Currently, images that use AI technology are labeled as ‘Made with AI’, encompassing both content generated from scratch using AI and images that have been touched up using AI tools.
Some professional photographers have voiced dissatisfaction with this labeling approach, arguing that their authentic content, slightly retouched using AI tools, was still labeled as ‘Made with AI’. Notable figures like former White House photographer Pete Souza were among those affected by this labeling inconsistency.
Meta acknowledged the feedback received and disclosed that they are working on a new labeling approach to indicate the extent of AI tool usage in labeled content. As part of this effort, they replaced the label ‘Made with AI’ with ‘AI Insights’ to provide users with more information on how AI was utilized in the image. The company updated its labeling statement for AI-generated content on its website to clarify that the labels did not always align with people’s expectations and lacked sufficient context.
Meta relies on industry-standard indicators used by other companies for labeling content created with AI tools to determine if something was made with AI. In cases where minor modifications were made using AI tools, such as adjusting colors or focus in an image, the content will now be labeled as ‘AI Information’. Users can click on the label to learn more about why it was labeled and the extent of AI usage in the image.