Using OpenAI’s API for detecting Explicit/Harmful/NSFW Content: Compliance and Account Safety

Dear OpenAI Community,

I am developing an application that aims to identify and classify explicit, harmful, or Not Safe For Work (NSFW) content. I am considering utilizing OpenAI’s API for this purpose and have the following questions:

  1. Is it permissible to use OpenAI’s API to detect and classify explicit, harmful, or NSFW content? Specifically, I intend to process user-generated content to identify material that violates community guidelines or is deemed inappropriate.

  2. Are there any risks of account suspension or banning associated with processing such content through OpenAI’s API? I want to ensure that my use case aligns with OpenAI’s usage policies and does not inadvertently lead to violations.

I have reviewed the OpenAI Usage Policies and understand that generating or promoting explicit or harmful content is prohibited. However, my objective is to detect and filter such content to maintain a safe environment for users.

I would appreciate guidance on best practices for implementing this functionality in compliance with OpenAI’s policies. Additionally, if there are recommended approaches or alternative solutions for content moderation tasks, please share them.

Thank you for your assistance.

Best regards,

A.

Although the accuracy may not be very high, a free moderation endpoint is available for a preliminary check to ensure compliance with the content policy.

By utilizing this moderation endpoint, potentially harmful content can be screened out.

https://platform.openai.com/docs/guides/moderation

1 Like