Weird content management policy

CodeKidz 25

Yesterday I came up with an error from Azure OpenAI Service. 'The response was filtered due to the prompt triggering Azure OpenAI's content management policy. Please modify your prompt and retry. To learn more about our content filtering policies please read our documentation: https://go.microsoft.com/fwlink/?linkid=2198766'

I checked my error console and found the prompt as the image shown below.

Screen Shot 2023-11-06 at 10.10.10

As you can see, though I don't understand German, but I also know what this sentence means, it doesn't mention any content that violates policies. My website is a programming education website for children, and users only tell me their names and ages. Why was it filtered?

Also, it seems like there's no issue when answering in other languages. I think OpenAI should look into this problem and check the questions in the training dataset.

And it only occurs in Microsoft OpenAI Service, when I use OpenAI's official API, the model is turbo-3.5-0613, this issue doesn't happen. I wonder whether there is any difference between the default prompt of these two APIs.

AshokPeddakotla-MSFT 33,101 Reputation points

2023-11-06T10:36:58.7866667+00:00

CodeKidz Greetings & Welcome to Microsoft Q&A forum!

'The response was filtered due to the prompt triggering Azure OpenAI's content management policy. Please modify your prompt and retry. To learn more about our content filtering policies please read our documentation: https://go.microsoft.com/fwlink/?linkid=2198766'

As per the error message, it seems that the response was filtered due to Azure OpenAI's content management policy. This policy is in place to ensure that the content generated by the AI model is appropriate and does not violate any policies. Did you check Content filtering categories already?

And it only occurs in Microsoft OpenAI Service, when I use OpenAI's official API, the model is turbo-3.5-0613, this issue doesn't happen. I wonder whether there is any difference between the default prompt of these two APIs.

Regarding your specific prompt, it's possible that certain words or phrases triggered the content management policy. Can you share more information about the prompt and the specific policies that were violated, it's difficult to say for sure why the response was filtered.

It's also worth noting that different AI models and APIs may have different default prompts and training datasets, which could explain why you're not encountering the same issue with OpenAI's official API.

Also, it seems like there's no issue when answering in other languages. I think OpenAI should look into this problem and check the questions in the training dataset.

The content filtering models have been specifically trained and tested on the following languages: English, German, Japanese, Spanish, French, Italian, Portuguese, and Chinese. However, the service can work in many other languages, but the quality might vary. In all cases, you should do your own testing to ensure that it works for your application.

In addition to the content filtering system, the Azure OpenAI Service performs monitoring to detect content and/or behaviors that suggest use of the service in a manner that might violate applicable product terms. For more information about understanding and mitigating risks associated with your application, see the Transparency Note for Azure OpenAI. For more information about how data is processed in connection with content filtering and abuse monitoring, see Data, privacy, and security for Azure OpenAI Service.

Do let us know if that helps or have any further queries.
CodeKidz 25 Reputation points

2023-11-06T11:53:44.7633333+00:00

@AshokPeddakotla-MSFT, Thanks for your patience in answering. I only received that error warning message, so I can't provide specific information on which policy was violated.

I have provide the prompts in the screen capture, which contains 1 system prompt, 1 assistant prompt and 1 user prompt. The user prompt is what caused the error, because I've tried combine with other historial messages. And It seems highly likely to happen again with Microsoft OpenAI Service. I use the same model gpt-3.5-turbo-0613 in Microsoft OpenAI.
CodeKidz 25 Reputation points

2023-11-11T04:17:29.8233333+00:00

Happened again, this time the answer is 'Sunniya 12 years old', I think maybe the age triggered this mechanism, but we are a children-oriented website, so mentioning age is unavoidable in the conversation.How can you help me?
CodeKidz 25 Reputation points

2023-11-11T04:30:38.46+00:00

Tried in playground, also been filtered, why does just mentioning age trigger a warning, it's so ridiculous.
Francesco Zuppichini (Contractor) 0 Reputation points

2024-05-03T12:36:35.94+00:00

this conversation made my day

Share via

Weird content management policy

Your answer