Anthropic, the developer behind the AI assistant Claude, has quietly updated its terms of service to reserve the right to request identification from subscribers. The change means that users of the paid version of Claude may be asked to provide a government-issued ID, such as a passport or driving licence, before continuing to use the service.
The policy shift is understood to be part of broader efforts to prevent misuse of AI tools, including fraud, impersonation, and the generation of harmful content. However, it has sparked concern among privacy advocates and businesses that rely on Claude for daily operations. The terms do not specify how often checks would occur or whether they would apply retrospectively.
For UK businesses, the implications are significant. Many companies have integrated Claude into customer service, content generation, and data analysis workflows. A sudden demand for ID could disrupt operations if employees or contractors are unable or unwilling to comply. Smaller firms, in particular, may lack the legal resources to assess the data handling risks associated with sharing sensitive identity documents with a US-based AI firm.
From a regulatory standpoint, the UK Information Commissioner's Office (ICO) has previously stated that any collection of personal data must be lawful, fair, and transparent. If Anthropic were to store or process ID documents, it would need a clear lawful basis under UK GDPR. Meanwhile, the EU AI Act, which may influence UK policy through the government's white paper on AI regulation, classifies certain identity management practices as high-risk, potentially requiring additional safeguards.
Dr. Eleanor Shaw, a technology policy researcher at the University of Cambridge, commented: 'This move by Anthropic is a double-edged sword. On one hand, it could reduce botnets and fake accounts. On the other, it creates a new vector for data breaches and raises questions about how long identity information is retained. For UK consumers, the risk is that their privacy is traded for security without clear benefit.'
For now, UK subscribers to Claude have not reported any widespread identity checks, but the terms are enforceable immediately. Users should review their account settings and consider whether they are comfortable with the potential requirement. The development also adds urgency to the UK government's ongoing consultation on AI safety and data protection, with consumer groups calling for clearer rules on when AI companies can demand personal identification.