Artificial intelligence research company Anthropic has announced the integration of invisible watermarking technology across content generated by its AI chatbot, Claude, enabling automated systems to identify AI-generated text and images.
The initiative aligns with transparency and content-attribution compliance mandates established under the European Union’s Artificial Intelligence Act. Anthropic confirmed that the watermarking protocol will be deployed globally and applied across specific usage environments, including Claude Code.
While invisible to the human eye, these embedded markers can be detected by specialized detection algorithms to confirm AI authorship. The decision comes amid growing industry concerns over the proliferation of low-quality, mass-generated AI content across the internet—commonly referred to as “AI slop.” However, tech experts note that while image watermarking is increasingly common, its efficacy can still be impacted by re-formatting, editing, or screenshotting.
