Microsoft will enable Anthropic’s Claude artificial intelligence models by default for UK Copilot customers from December 9.
The move means customer data processed through the models will be handled in the United States, according to a report by The Register.
The UK organisations that have not selected a preference by the deadline will have Claude automatically activated.
However, administrators can still disable the feature through their settings.
Leaving the setting unchanged will effectively amount to opting in to the use of Anthropic’s models.
According to Microsoft’s Copilot documentation, Anthropic operates as a Microsoft subprocessor, meaning it handles customer data under Microsoft’s contractual arrangements and oversight.
Claude is already enabled by default for most of Microsoft’s commercial customers, with the European Union, the European Free Trade Association and the United Kingdom previously excluded from the arrangement.
“This affirms Microsoft’s commitment to offering choice between leading AI models while maintaining enterprise-grade security and compliance,” Microsoft wrote.
Microsoft’s documentation also states that Claude models used in Copilot fall outside its EU Data Boundary and are not covered by its in-country data processing commitments. Since the UK is not covered by the EU Data Boundary, Microsoft is changing the default settings for British customers, The Register reported.
The change applies to Claude integration across Copilot, Researcher, Copilot Studio, Power Platform and Copilot within Microsoft 365 applications.
Microsoft is also discontinuing a separate setting that previously restricted the use of Claude models to Microsoft 365 applications.
