Speaker Recognition API returns "Unauthorized – This is a gated service" when creating a verification profile

Question

Speaker Recognition API returns "Unauthorized – This is a gated service" when creating a verification profile

Irshad Gouri 60

Hello,

I am trying to create a Text-Independent Speaker Verification Profile using the Azure Speaker Recognition API (Central India region). However, every request returns an Unauthorized error indicating that the API key is not approved for this gated service.

Here is the exact request I am using:

curl -X POST "https://centralindia.api.cognitive.microsoft.com/speaker-recognition/verification/text-independent/profiles?api-version=2021-09-05" \
  -H "Ocp-Apim-Subscription-Key: <my-key>" \
  -H "Content-Type: application/json" \
  -d '{"locale":"en-us"}'

And this is the response:

{
  "error": {
    "code": "Unauthorized",
    "message": "The API Key provided is not authorized. This is a gated service, make sure your Azure Subscription ID is approved: https://aka.ms/azure-speaker-recognition . For more information, please visit https://aka.ms/SpeakerRecoTransparencyNote ."
  }
}

I have verified that:

My Cognitive Services resource is successfully created in the Central India region.

The key and endpoint are correct.

The Speaker Recognition API version (2021-09-05) matches Microsoft documentation.

My question is:

Do I need to explicitly request access/approval for Speaker Recognition APIs for my Azure Subscription before these endpoints work? If yes, what is the exact process or expected approval timeline?

Any guidance or confirmation would be appreciated.

Thank you.

1 answer

Your answer

Answer 1

Hello Irshad,

Yes, based on the error message you're receiving, you absolutely need to explicitly request and receive approval for the Speaker Recognition API for your Azure Subscription before the endpoints will work.

The error message states: "This is a gated service, make sure your Azure Subscription ID is approved: https://aka.ms/azure-speaker-recognition". This indicates that access is not automatically granted when you create a Cognitive Services resource.

Here's a breakdown of the process and what to expect:

The Approval Process

Gated Access: Microsoft has placed the Speaker Recognition service, along with other sensitive AI services like Face API, under a "gated" access policy. This is to ensure responsible and ethical use of the technology, as outlined in their Responsible AI principles.
Request Form: To gain access, you must fill out a specific request form. The link provided in the error message (https://aka.ms/azure-speaker-recognition ) directs you to this form.
Required Information: In the form, you will need to provide details about your Azure Subscription ID and describe your intended use case for the Speaker Recognition API. Be prepared to explain why you need to use this service and how you will use it responsibly.
Review and Approval: After submitting the form, Microsoft's Responsible AI team will review your request. They are evaluating whether your use case aligns with their ethical guidelines and acceptable use policy.

Expected Approval Timeline

The approval timeline can vary. While there isn't a publicly guaranteed SLA, the process typically involves a review by a human team, so it is not instantaneous. You should expect to wait several business days, and in some cases, it could be longer depending on the volume of requests and the complexity of your use case. You will usually receive an email notification once a decision has been made.

In summary, your request is failing not because of a technical misconfiguration, but because of a policy-based access control. You must go through the approval process at https://aka.ms/azure-speaker-recognition to have the service enabled for your subscription.

Share via

Speaker Recognition API returns "Unauthorized – This is a gated service" when creating a verification profile

1 answer

The Approval Process

Expected Approval Timeline

Your answer