Meta's AI Images Now Have Invisible 'Digital Fingerprints'

Meta has introduced a preview of a web-based tool designed to identify images generated by its Muse Image AI model. This tool checks for an invisible watermark embedded by a system known as Content Seal. According to Meta, this watermark is resilient, remaining intact even after images are cropped, compressed, resized, or screenshotted.

Previously, Meta has released open-source versions of its watermarking technology. However, the Content Seal system integrated with Muse Image is proprietary. Unlike some earlier Meta AI versions that featured visible watermarks, the new approach relies on imperceptible digital signatures.

Initially, the detection capabilities are focused on images created or edited using Muse Image. Meta has stated its intention to expand the Content Seal watermarking to AI-generated and edited videos in the future. The company is also developing a separate video generation model, Muse Video, expected soon.

Testing of the detection feature confirmed its ability to identify watermarks on both fully AI-generated images and those edited with Meta AI, as well as in screenshots. A positive detection indicates the image was processed by Meta AI. However, the detection functionality is not yet integrated into the Meta AI app itself, and the web tool has limitations, including incompatibility with other watermarking standards and rate limits on daily checks.