Hi everyone,
I’m a paying ChatGPT Plus subscriber, but I’m currently unable to use Codex because OpenAI requires phone verification and its own fraud protection system is blocking my legitimate phone number.
The UI simply says:
“Unable to send a verification code to this phone number. Please try again later or use a different number.”
However, after checking the actual API response in the browser developer tools, the real error is:
code: "fraud_guard"
message: "We've detected suspicious behavior from phone numbers similar to yours. Please try again later or contact us through our help center."
So this is not an SMS delivery problem. The verification request is being actively rejected by OpenAI’s fraud_guard system before a verification code can even be sent.
I contacted OpenAI Support and clearly explained this issue with a screenshot showing the API response. Unfortunately, the response I received did not address the actual problem. I was told that phone verification cannot be manually bypassed, reset, or changed, and that I should either recover my original phone number or create a new OpenAI account with another phone number.
But I do have full access to my original phone number. It is a legitimate Vietnamese mobile number that belongs to me. I am not asking to bypass verification. I am reporting what appears to be a false positive in OpenAI’s fraud detection system.
What makes this particularly frustrating is that I am already paying for ChatGPT Plus, yet I cannot access a service available to my account because my legitimate phone number has been automatically flagged, and Support appears unable or unwilling to investigate the actual fraud_guard error.
I would like OpenAI to clarify:
- How can a legitimate user appeal a false-positive
fraud_guardblock? - Is there any way to escalate this issue to the technical or fraud-prevention team?
- Why should a paying Plus subscriber be told to create an entirely new account instead of having an incorrectly flagged phone number properly reviewed?
Has anyone else experienced the same fraud_guard error with a legitimate phone number? If so, were you able to get it resolved?
I can provide screenshots showing both the user-facing error and the actual API response.
