June 23, 2026, (Inside AI) — Anthropic will require identity verification for a small subset of Claude users starting July 8, using third-party service Persona to verify government-issued photo IDs and live selfies. The policy targets accounts flagged for potential violations, offering a path to appeal rather than immediate restriction.
The move aims to bolster platform safety, curb misuse, and satisfy evolving legal demands. Yet it also surfaces tensions between security and privacy, amplified by Persona’s investor ties to Anthropic and the collection of facial geometry data under state biometric laws.
Anthropic said the checks will not apply universally. Instead, a limited number of users whose accounts trigger internal flags will be prompted to complete verification. Those who comply can regain access to certain features; those who do not may face continued restrictions.
The process, which takes under five minutes, requires a valid passport, driver’s license, or national identity card. Photocopies, screenshots, mobile IDs, and expired documents are rejected. Users must also submit a live selfie via phone or webcam for liveness detection.
Privacy protections are layered but not absolute. Persona stores the ID and selfie data on its own servers, not Anthropic’s. However, Anthropic can access verification records through Persona when necessary, such as during an appeal. The company stressed it does not use verification data for model training, and all data is encrypted in transit and at rest.
Critics have raised red flags. Persona is backed by Founders Fund, a venture firm that also invests in Anthropic, prompting questions about independence. More pointedly, the live selfie generates facial geometry data, which may fall under strict biometric privacy laws in states like Illinois and Texas. Anthropic has not detailed how it navigates these legal frameworks.
If verification fails, users can retry with better lighting or a different ID. Those who believe their accounts were wrongly restricted can appeal through Anthropic’s support process, though the company did not specify the timeline or success rate for such appeals.
The policy mirrors a broader industry shift toward identity gating for AI platforms, as regulators worldwide push for stronger user accountability. Anthropic’s limited rollout suggests a cautious test of the waters, but the privacy implications and investor overlap will likely draw continued scrutiny.
