Anthropic's updated privacy policy gathers biometric data from users identified as flagged on Claude.

      TL;DR: Anthropic's revised privacy policy mandates that some Claude users must provide government IDs and selfies for identity verification.

      Anthropic has updated its privacy policy, allowing the company to require certain Claude users to upload government-issued identification and submit selfies or videos for identity verification. The new policy, effective July 8, introduces a category for collecting personal data, which includes facial geometry templates that could be classified as biometric information under state privacy regulations.

      TechCrunch first reported on the change, detailing the updated identity verification policy released on June 17. Anthropic spokesperson Thariq Shihipar mentioned that this requirement applies to a “small subset of users” whose accounts have been flagged for potential policy violations, providing a means to appeal instead of facing complete account suspension.

      Accepted identification includes passports, driver’s licenses, state or provincial IDs, and national identity cards. Digital IDs, screenshots, and photocopies are not permissible. Users submitting verification materials will also need to provide a selfie or video, which the system uses to create a facial geometry template for comparison with the provided document.

      Anthropic does not carry out the verification in-house and instead utilizes Persona, a San Francisco-based identity verification platform, to process the documents and biometric data. Persona is supported by Founders Fund, the venture firm led by Peter Thiel, who is also an investor in Anthropic through that same fund.

      This connection has faced scrutiny previously. In February 2026, Discord chose Persona for its age verification system but reversed its decision after user backlash regarding the sharing of government IDs with a company associated with Thiel. Additionally, a separate security incident exposed Persona data on a US government-authorized endpoint, revealing approximately 2,500 accessible files, though the extent and sensitivity of those files remain unclear.

      The collection of facial geometry data by Anthropic raises legal concerns, as Illinois’s Biometric Information Privacy Act designates facial geometry as biometric data, imposing fines of $1,000 to $5,000 for companies collecting it without proper consent. Facebook settled a class action under BIPA for $650 million in 2021, establishing a precedent for biometric privacy litigation in the US.

      Shihipar noted that the identity verification policy is “unrelated to the Fable or Mythos rollout,” which refers to the Trump administration's mandate that required Anthropic to disable its most potent AI models earlier this month. However, the timing is difficult to disentangle from the growing tensions between the company and Washington, which have included a Pentagon supply-chain risk designation, the shutdown of Fable 5 and Mythos 5 due to export-control issues, and concerns raised by Amazon CEO Andy Jassy about Anthropic’s models with the White House.

      Anthropic boasts tens of millions of monthly users, insisting that the verification requirement will impact only a small portion of them. However, for a company that has built its reputation on safe and responsible AI deployment, asking users to provide passport scans and facial biometric data to a Thiel-affiliated vendor raises trust issues that technical assurances may not fully address.

      Last week, President Trump indicated that he no longer perceives Anthropic as a national security threat after a G7 meeting with CEO Dario Amodei. Whether this easing of tensions translates to alleviated regulatory pressures that may have prompted stricter user verification remains uncertain.