Will It Become Easier to Distinguish Between AI and Human-Generated Content?
AI company Anthropic announced a new global initiative to enhance the detection and transparency of content generated by its AI assistant, Claude. The company stated that it will comply with the transparency and labeling requirements established under the European Union’s AI Act. Accordingly, Anthropic is rolling out invisible digital watermarking for text, code, and other Claude-generated outputs to users worldwide.
With this move, Anthropic joins the ranks of technology companies that turn regional compliance requirements into global standards rather than scoping them solely to specific jurisdictions. The technology embeds machine-readable, imperceptible digital signals directly into Claude’s outputs that remain invisible to the human eye while being detectable by automated algorithms.
The decision comes under Article 50 of the EU AI Act, which mandates the machine-readable labeling of generative AI outputs. Instead of restricting these safeguards geographically, Anthropic integrated them directly across its entire system. The rollout applies across Claude’s consumer applications, developer APIs (Claude Platform), tools such as Claude Code, Claude Cowork, and Claude Tag, as well as cloud deployments via AWS, Google Cloud, and Microsoft Foundry. This move is expected to accelerate similar steps from competitors like Google and OpenAI.
This worldwide implementation is seen as a critical step in curbing AI-driven disinformation, copyright infringements, and academic misconduct. Enabling technical verification of content origin will allow platforms and regulatory authorities to reliably distinguish AI-generated material from human-authored work.