Unexpected Behavior When Using openai.ChatCompletion.create() with Custom user ID

[Anipaleja]

Confirm this is an issue with the Python library and not an underlying OpenAI API

• This is an issue with the Python library

Describe the bug

When passing a custom user value to openai.ChatCompletion.create(), the request completes successfully, but there is no indication that the value is being acknowledged, logged, or used in any visible monitoring system. The documentation states that the user field is used for abuse monitoring, but from the client perspective, it is impossible to confirm whether this information is being handled or ignored.

This makes it difficult to audit usage or confirm compliance with OpenAI’s best practices.

To Reproduce

• Use the latest version of openai Python package.
• Run the code:

import openai

openai.api_key = "your-key-here"

response = openai.ChatCompletion.create(
    model="gpt-4",
    messages=[{"role": "user", "content": "Hello"}],
    user="my_test_user_001"
)
print(response)

• Observe that the request works, but there is no way to verify the user field was registered or tracked.

Code snippets

import openai

openai.api_key = "api_key"

response = openai.ChatCompletion.create(
    model="gpt-4",
    messages=[
        {"role": "user", "content": "Hello!"}
    ],
    user="test_user_123"
)

print(response)

#This snippet Demonstrates correct usage of the user parameter and shows the lack of any visible effect or confirmation that the user ID was acknowledged in the response.

OS

macOS

Python version

Python v3.11.3

Library version

openai v1.30.1

[Vishal-sys-code]

Hi @Anipaleja,

Thanks for raising this important point about the user parameter in Chat Completions. I've looked into this.

My investigation confirms the following:

1. The openai-python SDK correctly sends the user string you provide in client.chat.completions.create(..., user="your_user_id") to the OpenAI API as part of the request.
2. The OpenAI API documentation (and the SDK's docstrings) state that this user parameter is primarily used by OpenAI for abuse detection and monitoring.
3. Currently, the OpenAI API does not include a field in the response object that explicitly confirms the reception or processing of the user ID. This means that while the ID is sent, you won't see it echoed back in the completion response. This is the current design of the API.

To help clarify this for future users of the SDK, the docstrings for the user parameter in the Python library have been recently updated to note that the API does not return a direct confirmation for this value.

Since this behavior is inherent to the OpenAI API, any change to provide such a confirmation would need to come from OpenAI. If you'd find this feature valuable, you might consider providing this feedback directly to OpenAI through their official channels.

P.S. I also noticed the code snippet in the original report (openai.ChatCompletion.create(...)) uses syntax from an older version of this library (pre-v1.0.0). For users on openai version 1.0.0 or newer, the correct syntax is:

from openai import OpenAI
client = OpenAI() # Assumes OPENAI_API_KEY is set
response = client.chat.completions.create(
    model="gpt-4",
    messages=[{"role": "user", "content": "Hello"}],
    user="test_user_123"
)
print(response)

You can check the Migration Guide here: #742. This will help you a lot for the other syntaxes too...

Thank you....

[RobertCraigie]

This is the old syntax, please update to v1.