Artificial Intelligence: New watermarks to make AI texts recognizable
The company Anthropic, a well-known developer of artificial intelligence (AI), has announced an important innovation for its language models. In the future, texts created by the AI "Claude" will be marked with a digital watermark. This technology is intended to help better distinguish machine-generated content from human-written texts.
The system works through a special, machine-readable pattern that is embedded into the text. This pattern is not visible to users, but special programs can identify it. Anthropic also plans to introduce a paid interface, a so-called detection API, which companies and organizations can use to check whether a text was written by an AI.
This development is particularly significant for the education sector and the media world. In Austria, there is currently an intensive debate about how AI tools can be used safely in schools and universities. The new technology could help teachers and employers make the origin of texts more transparent and reduce the misuse of AI-generated content.
Experts see this measure as an important step toward greater security and trust in the digital space. Nevertheless, the question remains as to how reliably such detection systems will function in the long term, as technology for bypassing such watermarks is also constantly being further developed.
In summary, the labeling of AI content remains an important topic for the digital future. Whether Anthropic's new watermarks will establish themselves as a standard will become clear in the coming months as the technology becomes more widely available.