Anthropic has begun embedding invisible, machine-readable watermarks into text generated by Claude, creating a hidden signal that can travel with AI-written content even after it is copied and may survive some forms of editing.

Anthropic disclosed this on Monday, saying its marking system is part of its commitments under the European Union’s AI Act Code of Practice on Transparency of AI-Generated Content.

The company said Claude models launched on or after August 2, 2026, support machine-readable marking from launch, while it is working to add the capability to models released before that date.

Anthropic said the move is intended to provide greater transparency as AI-generated content becomes increasingly common.

“As AI-generated content becomes commonplace, greater transparency and signals about where content comes from can give people useful context about the information they consume,” the company said.

For text, Claude embeds the watermark directly into the generated output.

“When a supported Claude model generates text, it weaves an imperceptible watermark directly into the text itself,” Anthropic said, adding that the mark does not change the “meaning, quality, or readability” of the response.

The company said the mark can travel beyond Claude when users copy the text elsewhere.

“Because the watermark is part of the text, it will travel with the text when it’s copied and pasted elsewhere, and may persist through some editing,” Anthropic said.

Anthropic is also developing systems that will allow users and third parties to detect Claude’s marks, although it said detailed information about the detection mechanisms will be provided in forthcoming technical documentation.

However, Anthropic warns that detecting a Claude mark is not conclusive proof that Claude originally authored the content. This is because people often use Claude to proofread, translate, summarize, or convert files.

Conversely, the absence of a detectable Claude mark does not prove that a piece of work was created entirely by a human.

Anthropic said marks can become undetectable if text is heavily edited, paraphrased, translated or mixed with other writing. Very short passages may also contain insufficient material for reliable detection.

The development comes as the rapid adoption of generative AI is creating growing questions about how much of the work being submitted, published or consumed online is actually produced by humans.

The widespread adoption has raised concerns about whether AI is being used to improve people’s abilities or to replace the effort required to develop them.

In February 2025 that Nigerian university lecturers were raising concerns over students increasingly turning to AI tools such as ChatGPT to complete assignments and answer tests.

Prof. Isaac Nwaogwugwu of the University of Lagos told Nairametrics that about 40% of more than 100 MBA students in one of his classes submitted identical answers after apparently using the same AI tool.

The concerns have also extended beyond students. Dr. Ereke of Ebonyi State University told Nairametrics that some lecturers were themselves ‘lazily’ using AI to generate lesson notes, course outlines, marking schemes and examination questions without adequately reviewing them.

This creates a particularly important context for AI watermarking.

If students can use generative AI to produce assignments that appear to be their own work, a machine-readable signal embedded within AI-generated text could eventually give educational institutions another way to identify whether submitted material has been processed by an AI system.

The Nigerian government has also begun incorporating artificial intelligence into Nigeria’s education system, with the technology being positioned as a tool to improve teaching and learning while preserving human agency.

The Minister of Education, Tunji Alausa, disclosed this during a ministerial press briefing in Abuja to mark the International Day of Education, themed “AI and Education: Preserving Human Agency in a World of Automation.”

Represented by Dr. Ejeh Usman, Director of Polytechnics in the ministry, Alausa said the ministry had integrated AI-powered platforms that provide comprehensive information on tertiary institutions and admissions across the country as part of efforts to incorporate AI into teaching and learning.