Why ChatGPT and Claude Are Adding Facial Real‑Name Verification

ChatGPT recently displayed a pop‑up that requires users to upload a government‑issued photo ID and record a live selfie for 3D liveness detection.

OpenAI age‑prediction model

On January 20 2026 OpenAI silently launched an age‑prediction model that runs without user notice. The model cross‑references four signals:

How long the account has been registered

Time windows of highest activity

Typical usage patterns

The age the user entered in their profile

These signals are combined to produce a probability that the user is a minor. When the probability exceeds a threshold, ChatGPT automatically blocks content deemed potentially harmful for minors. The restriction is applied silently; users receive no notification. To remove the restriction, users must complete identity verification, which may involve a live selfie, a government‑issued ID, or both, depending on the region.

Developer‑side organization verification

In April 2025 OpenAI extended verification to API developers. Access to advanced models requires uploading a physical document and passing a “Verify Organization” check. Only physical documents are accepted; digital IDs, driver’s licenses, and other electronic credentials are rejected. Each document can be used to verify only one organization within a 90‑day window, effectively preventing “sock‑puppet” accounts. Organizations that fail verification are denied access to advanced models.

Anthropic identity verification

On June 10 2026 Anthropic emailed Claude users informing them that, starting July 8, the updated privacy policy will enforce identity and age verification. Verification requires an official photo ID and a camera‑enabled device to capture a live selfie, mirroring OpenAI’s approach.

Motivations

Abuse prevention – verification aims to stop large‑scale misuse such as bulk account creation, illicit fine‑tuning, or model scraping.

Agent accountability – as AI agents perform real‑world actions (e.g., booking flights, editing documents), knowing who issued the command becomes a legal and safety requirement.

Timeline

Jan 2026 – OpenAI silently launches the age‑prediction model.

Apr 2025 – OpenAI requires organization verification for API developers.

Jun 2026 – Anthropic announces upcoming verification for Claude users.

Jun 2026 – ChatGPT consumer‑side facial verification pop‑up appears.

References

https://help.openai.com/en/articles/12652064-age-prediction-in-chatgpt

https://help.openai.com/en/articles/8774323-why-am-i-being-asked-to-verify-my-identity

https://help.openai.com/en/articles/10910291-api-organization-verification

https://help.openai.com/en/articles/10362446-api-reasoning-model-availability-by-usage-tier

Code example

关注我
们,
设为星标,每天7:40不见不散
回复
架构师
获取资源
大家好，我是你们的朋友架构君，一个会写代码吟诗的架构师。
来源：新智元