Table of Contents
ToggleAnthropic quietly launched an identity verification documentation on the Claude support page yesterday (14th), officially announcing that some Claude features will require users to complete real-name verification. The official statement said: "Use powerful technology responsibly, starting with knowing who is using it."
The purpose of identity verification covers three aspects: preventing abuse, enforcing usage policies, and complying with legal obligations.
This policy is currently only applicable in certain situations and is not mandatory across the board. Users may encounter verification prompts in the following situations: when accessing specific advanced functions, when the system performs routine platform integrity checks, or when security and compliance measures are triggered.
Verification process: Passport or driver's license
According to the official instructions, the following conditions must be met to complete the verification:
A valid government-issued photo ID (physical file, not a photocopy or screenshot), and a mobile phone or computer with a camera to complete the process.
Acceptable document types include passports, driver's licenses, and national identity cards. Explicitly unacceptable documents include: photocopies, screenshots, digital or mobile versions of documents, non-government-issued documents, and temporary paper documents.
Note: Anthropics has selected Persona Identities as its verification technology partner.
How is data protected? Anthropic declares it will not touch original data.
Anthropic explicitly states that it is the "data controller," but the actual data processing is carried out by Persona. Furthermore, the Anthropic system itself does not hold users' ID images or selfies; this data is collected and stored by Persona.
Anthropics has contractual restrictions on the use of Persona, limiting it to authentication and fraud prevention purposes only, and prohibiting its use for other purposes. All data is encrypted during transmission and storage at rest.
Furthermore, the official statement specifically emphasized two negative announcements, explicitly excluding the use of data for model training and sharing it with third parties for marketing or advertising. Anthropic stated that it only collects the minimum necessary information.
What if verification fails? Account ban conditions are also disclosed.
If the verification process is not smooth, users can try again multiple times or contact the support team through the form.
It's worth noting that Anthropic has also publicly disclosed the triggering conditions for account bans, including: violating usage policies, originating from unsupported regions, violating terms of service, and being under 18 years of age. Banned users can appeal through the appeal form.
This is inevitable under compliance pressure, but the increased threshold raises concerns.
It's not hard to guess that this policy comes against the backdrop of increasingly stringent global compliance requirements faced by AI service providers.
For Anthropic, while accelerating its expansion into the enterprise (Claude to B) market, establishing a verifiable user identity foundation is essential to attracting clients in highly regulated industries such as finance, healthcare, and law.
However, from the user's perspective, groups that originally used Claude anonymously or in a low-friction manner, such as researchers and users with strong privacy concerns, may have concerns about this.
Although the official statement emphasizes that the data is not stored on the Anthropic system, it remains to be seen whether the "third-party custody" structure can truly eliminate privacy concerns.
The current verification mechanism's triggering conditions are not yet fully transparent. Which "specific functions" require verification and whether its scope of application will be gradually expanded are key questions worth tracking. If this mechanism is linked to subscription plans or API usage in the future, its impact on the developer community will be even more direct.
