Anthropic Limits Claude's Content Privacy
Anthropic announced the addition of hidden digital watermarks to content generated by the AI assistant "Claude," aimed at enhancing transparency and enabling source verification. The technology involves embedding imperceptible marks into texts and images, utilizing standards such as C2PA for images, along with support for digital signatures and metadata that remain associated with the content when it is copied or modified. This move aligns with the requirements of the European AI Act, which began its enforcement in August 2026, with developers given a grace period of up to four months to comply. The technology also supports services like AWS, Google Cloud, and Microsoft Foundry, and includes tools developed to aid in detecting these marks, although their effectiveness is not always guaranteed. The company's goal is to ensure that AI-generated content can be more accurately traced and verified for its origin, while acknowledging that some information may be deleted or concealed in certain cases.
Notice: This Is an AI-Generated Summary
Comments (0)