SAN FRANCISCO — Meta has launched a proprietary invisible watermarking system designed to identify and track images generated by its artificial intelligence models. The technology, known as Content Seal, represents the company's latest effort to address growing concerns regarding the spread of deceptive synthetic media across its digital platforms.
According to reports from other news platforms, the deployment follows a directive issued in March by the company's independent Oversight Board. The board had publicly urged the social media operator to honor its safety commitments and implement internal tools capable of detecting and labeling artificial intelligence outputs.
The new tool embeds imperceptible markers into images produced by the company's generative models, allowing automated systems to verify their origin. However, reporters around the world have noted that the decision to build a proprietary framework has drawn scrutiny from some industry observers, who argue that utilizing established external standards could have provided a more integrated solution.
The debate highlights an ongoing division within the technology sector over whether to adopt unified, cross-industry detection standards or rely on individual corporate systems. Critics suggest that independent, siloed watermarks may struggle to maintain efficacy when content is transferred across different social networks and digital environments.
While the company continues to monitor the rollout of its new detection mechanism, independent researchers are expected to evaluate how resilient the invisible markers are against attempts to alter or strip metadata. The implementation marks a significant step in the broader corporate effort to mitigate the influence of synthetic misinformation online.
