Anthropic just added a new safeguard that's raising eyebrows: identity verification. Some Claude users are now being asked to provide a passport, driver's license, or government ID, along with a live selfie. The company says it's rolling this out for "a few use cases," according to its Help Center.
Anthropic calls itself the "data controller," setting the rules for how and where your ID data is used and stored. But the actual collection and storage is handled by Persona Identities, an ID verification startup.
Persona is contractually obligated to use that data "only to provide and support verification and to improve their ability to prevent fraud," Anthropic said.
So why is an AI chatbot asking for your driver's license?
"This applies to a small number of cases where we see activity that indicates potentially fraudulent or abusive behavior, which violates our usage policy," an Anthropic spokesperson told Business Insider.
If Anthropic decides the activity breaks its rules, the user's account could be banned.
Anthropic's help page lists the potential reasons an account might get banned even after ID verification:
- Repeated violations of our Usage Policy
- Account creation from an unsupported location
- Terms of Service violations
- Under-18 usage
There's also an appeals form for users who think they've been wrongfully banned.
Claude users aren't thrilled about the ID checks
Users on X have already started noticing the requests. One posted a screenshot of a prompt in Claude asking for a "quick identity check," claiming it would only take two minutes and required an ID and mobile camera access. (See: Anthropic Could Raise Billions at $800B Valuation, But It's Not Biting) When AI Models Promise Security, They Might Be Creating a Whole New Problem) Claude Code Token-Saving Guide: Why Frequent Restarts Cost More Than You Think)
https://twitter.com/hqmank/status/2044332406417830255
Another screenshot shows the completion screen: "Thank you for verifying your identity," accompanied by a celebratory graphic.
The backlash was swift. "Anthropic making unexplainable decisions," one user wrote. "We are living in 1984," another added.
https://twitter.com/xw33bttv/status/2044334621849223588
https://twitter.com/zamdoteth/status/2044724923730419951
In its Help Center, Anthropic also listed what it's not doing with this data. The company says it's not training its AI models on the ID verification data, and it's not sharing that data with anyone beyond Anthropic and Persona, except where legally required.
"We are not collecting more than we need," Anthropic wrote. "We ask for the minimum information required to verify your identity."