Anthropic is tightening user verification for its Claude AI service. Beginning July 8, the company will ask certain users whose accounts have been flagged for potential policy breaches to submit a government‑issued ID—such as a passport, driver’s license, state or provincial ID, or national identity card—along with a selfie photo or video. The selfie will be used to generate a facial geometry template, a data point that state privacy statutes, including Illinois’ Biometric Information Privacy Act, treat as biometric information.
The policy applies only to a "small subset of users," according to Anthropic spokesperson Thariq Shihipar. Those users have not been banned outright; instead, the verification step offers a route to appeal the flagging decision. Digital copies, screenshots, or photocopies of IDs will not be accepted.
Anthropic does not process the documents or biometric data in‑house. The company has contracted Persona, a San Francisco identity‑checking platform, to handle the verification workflow. Persona, which counts Founders Fund—led by Peter Thiel—as a backer, has faced scrutiny after Discord’s 2026 age‑verification rollout drew user backlash over sharing government IDs with a Thiel‑affiliated firm. A separate incident revealed Persona data on a U.S. government‑authorized endpoint, though the extent of the exposure remains unclear.
Collecting facial geometry raises legal questions. Illinois’ Biometric Information Privacy Act (BIPA) classifies such data as biometric and imposes penalties of $1,000 to $5,000 per violation when companies gather it without proper consent. The law has been the basis for high‑profile settlements, including Facebook’s $650 million class‑action resolution in 2021. Anthropic’s move therefore places the company under the same legal microscope that has affected other tech firms.
While Anthropic emphasizes that the new verification requirement is unrelated to recent regulatory actions—such as the Trump administration’s order to disable its most powerful AI models—the timing coincides with heightened scrutiny from Washington. The company has faced a Pentagon supply‑chain risk designation, export‑control shutdowns of its Fable 5 and Mythos 5 models, and reports that Amazon CEO Andy Jassy raised security concerns directly with the White House.
Anthropic serves tens of millions of monthly users, and the firm asserts that only a minor fraction will encounter the ID and selfie demand. Nonetheless, the requirement introduces a trust question for a company that has marketed itself as a leader in safe and responsible AI deployment. Users must now decide whether to provide passport scans and facial biometrics to a third‑party vendor tied to a prominent venture fund.
President Trump recently signaled that Anthropic is no longer viewed as a national‑security threat, following a G7 meeting with CEO Dario Amodei. Whether that diplomatic shift will ease regulatory pressure or influence the company’s verification policy remains uncertain.
Cet article a été rédigé avec l'assistance de l'IA.
News Factory APP - actualités agentiques pour booster votre SEO et AEO.