Anthropic is adding a new layer of user verification to its Claude AI platform. The change, outlined in a privacy‑policy update set to take effect on July 8, will require a small subset of users whose accounts are flagged for potential fraud or abuse to submit a government‑issued photo ID and a selfie. The company says the step lets those users appeal a suspension rather than be banned outright.

While Anthropic has long required users to be at least 18, the latest policy expands verification beyond age checks. When triggered, users must upload a scan of a passport or driver’s license, along with a selfie or video. Anthropic will also generate a digitized face‑geometry template, data that some states, such as Illinois, treat as protected biometric information. The firm will keep a record of the verification outcome, indicating whether the user meets the age requirement or has passed identity checks.

Anthropic’s spokesperson, Michael Aciman, pointed journalists to an X post by the company’s Thariq Shihipar, noting the update applies only to a “small subset of users” whose accounts are flagged but not outright banned. The firm did not disclose how many users fall into that category, though Anthropic reportedly serves tens of millions of monthly users.

The verification requirement is tied to Anthropic’s broader effort to navigate a complicated regulatory landscape. The company has been at odds with the White House, especially after Trump officials pressed the firm to withdraw its latest cybersecurity models amid concerns that a jailbreak could bypass safety guardrails. Earlier this year, the Department of Defense labeled Anthropic a “supply chain risk,” a move Anthropic viewed as retaliation for refusing to let the government use its technology for mass domestic surveillance or autonomous weapons.

By tightening its identity checks, Anthropic hopes to demonstrate compliance with a patchwork of state and federal regulations, as well as to address pressure from the current administration. The policy states that verification may be required “in certain circumstances,” though it does not spell out specific triggers.

Anthropic will rely on San Francisco‑based Persona to handle the ID checks. Persona will present a verification prompt when users access certain capabilities, as part of routine platform‑integrity and safety measures. While Anthropic decides how long Persona retains the uploaded documents, the company did not say when the data will be deleted. Persona, however, has faced scrutiny for its ties to Founders Fund, an investment firm founded by Peter Thiel, a known Trump backer. The firm also backs Anthropic.

Persona’s data‑retention practices have drawn criticism elsewhere. Roblox, another Persona client, says user images are deleted “immediately” after processing to limit exposure. Persona can still be compelled to turn over stored information to U.S. authorities, a risk that raises privacy concerns for users.

Earlier in the year, Discord briefly adopted Persona for age verification before reversing the decision after user backlash. The controversy highlights the tension between ensuring platform safety and protecting user privacy.

Anthropic’s updated privacy policy does not mention any change to its core services or pricing. Instead, it frames the identity‑verification step as a tool for “preventing and investigating fraud, abuse, and violations of its terms, including unlawful or criminal conduct,” and for “investigating and resolving security issues.”

As the policy rolls out, users will encounter verification prompts only when the system flags their accounts for review. The company says the process is designed to be routine and to support a smoother appeals pathway, rather than to create barriers for the majority of Claude users.

Cet article a été rédigé avec l'assistance de l'IA.
News Factory APP - actualités agentiques pour booster votre SEO et AEO.