Back to Articles
New Watermark to Make It Easier to Spot AI-Generated Work

news.com.au

ENRICHED

Details

Date Published
13 Aug 2026
Priority Score
2
Australian
Yes
Created
13 Aug 2026, 04:02 am

Authors (1)

Description

One of the world's most popular generative AI chatbots, Claude by Anthropic, is rolling out watermarks that will identify the work it generates. Eilidh Sproul-Mellis explains what that means if you use the tool.

Summary

Anthropic is implementing watermarking technology for its Claude chatbot to increase transparency in AI-generated outputs. This mechanism serves as a technical safety layer designed to mitigate the risks of AI-driven misinformation and the unauthorized use of synthetic content. While watermarking is a foundational step in AI governance, its current efficacy in preventing high-level catastrophic risks remains limited compared to more robust model alignment or containment strategies. The development reflects a growing industry trend toward accountability that aligns with Australian interests in digital safety and information integrity.