Search
In July 2026, Meta launched its AI content detection system, Content Seal, which embeds invisible watermarks into images generated by AI models such as Muse to distinguish between real and deepfake content. However, the system has raised questions. Tests by Reuters found that the detection tool could not identify the AI origin of images that had been cropped. Meta responded by stating that the tool is still in its preview stage and that significant cropping could result in the loss of the watermark signal. Additionally, despite being a member of the Content Provenance and Authenticity Alliance, Meta chose to develop its own solution instead of adopting existing mature options, leading to accusations of reinventing the wheel, which may reduce ecosystem responsiveness and external compatibility.
-
C114 Communication Network -
Communication Home
