Claude and ChatGPT users report account bans in large numbers, as both companies tighten identity and regional restrictions
Recently, many Claude and ChatGPT users have reported that their accounts were suddenly banned. Claude users have frequently run into identity verification requirements, and some were banned immediately after submitting a Chinese passport. ChatGPT, meanwhile, began seeing a wave of account-ban complaints around November 7, with some free users, Plus users, Pro users paying $200 per month, and even enterprise accounts affected.
Claude’s identity verification has been in place since April this year. Users may be asked to upload a passport or other documents through the third-party platform Persona, and then take a selfie to verify their identity. Since the end of October, many long-time users have suddenly encountered verification requests. Some were banned immediately after using a Chinese passport for verification, while others passed successfully. One user was banned instantly after submitting a passport, and their account was restored the next day after an appeal.
Anthropic officially prohibits use of Claude from unsupported regions such as mainland China. In its safety report in September, the company explicitly said that the system identifies risk signals such as unauthorized resale of Claude and accounts from restricted regions including China, and requires the relevant users to verify their identity; if verification fails, the account will be banned.
OpenAI has also begun strictly restricting multi-account usage. According to the company’s latest official documents, if too many subscriptions are linked to the same bank card, new accounts will be unable to purchase memberships; if too many paid accounts log in on the same device, restrictions will also be triggered, and it may even result in payment succeeding but the service being unusable. The restrictions cover Go, Plus, and Pro memberships at all tiers.
OpenAI recently also gave eligible Pro users API credits equal to the price of their membership, up to $500 per month. These credits have also become items for resale by third-party intermediaries for other users to use. OpenAI explicitly prohibits reselling API credits or sharing accounts, and violations may result in account bans.
The two companies have also successively updated their usage policies recently. Anthropic announced new rules on October 8, further clarifying access restrictions for unsupported regions such as mainland China, which took effect on November 12. OpenAI announced new rules on October 21, requiring intermediary platforms that directly provide access to GPT models to obtain official approval in advance, taking effect on November 20.
Original source:
Title: Countering misuse of AI: September 2026 Publisher: Anthropic Publication date: 2026-09-10 URL：https://www.anthropic.com/threat-intelligence-report-september-2026 Source nature: Company official threat intelligence report Verification status: This report’s relevant sections were actually read in this instance, verification date 2026-11-10 Supported content: Supports the policy statement that accounts from unsupported regions such as China and unauthorized resale of Claude are risk signals that require identity verification, and that failure leads to account bans
ABAB AI Insight
Since April, Anthropic has been using Persona to carry out identity verification for platform integrity checks and compliance. In its September threat intelligence report, it treated identity verification as one of the measures to counter distillation attacks and access from unsupported regions, explicitly naming accounts from China, Russia, and Iran. The report documented that Chinese labs such as Alibaba and DeepSeek carried out large-scale distillation through tens of thousands of fake accounts, which provided the background for the subsequent tightening of verification.
On the OpenAI side, the account-ban complaints in early November appeared alongside the previously announced device- and bank-card-based limits on multiple subscriptions. The limits target Go, Plus, and all Pro tiers, and are intended to curb multiple paid accounts linked to a single device or card. At the same time, OpenAI provided eligible Pro users with API credits of up to $500 per month, equal to the subscription tier value; these credits became targets for resale by intermediaries, while the company explicitly prohibits reselling API credits or sharing accounts.
As for policy updates, Anthropic’s new usage policy took effect on November 12, further clarifying usage restrictions for unsupported regions, including mainland China, covering a person’s physical location, the entity’s place of registration, and majority ownership control. OpenAI updated its policy on October 21, requiring intermediary services such as model routers and AI gateways that directly provide model access to obtain approval in advance, with effect from November 20, and it made explicit rules such as banning the use of outputs to train competing models.
Together, these measures target regional circumvention, account sharing, credit resale, and model distillation. For users, the risk of accessing unsupported regions or bypassing restrictions with a VPN has increased; for intermediary platforms, approval must be completed or the service model adjusted before the effective date. Indicators to watch include the success rate of account-ban appeals, the frequency of verification triggers, and whether intermediary platforms publicly disclose approval before November 20.
ABAB News · Cognitive Law
Anthropic uses identity verification as a tool to identify accounts from unsupported regions and unauthorized resale, and verification failure can directly lead to account bans.
OpenAI uses device and bank-card restrictions to curb multiple paid accounts, while also prohibiting API credit resale and account sharing.
The two companies have successively updated their policies, clarifying approval requirements for intermediary platforms and access restrictions for unsupported regions.