Questions on Usage Guidelines

Hi there,

I’m currently building a web application using OpenAI and I’m wondering about a few rules in your ToS/Usage Guidelines.

Do not permit end-users to input custom prompts (i.e., end-users should not have a Playground-like interface where they choose the general goal of the prompt)

So I’m not allowed to give the users the ability to craft their own prompts? For example Write a title for a blog post about artificial intelligence. Instead, I need to prompt-engineer it in a way where the user only enters artificial intelligence and the output would be a blog title for artificial intelligence.

Did I understand this correctly?

If so, does this mean I cannot implement a long-form feature like Google Docs where users can enter any prompt they wish?

Do not permit automated execution of model-generated code.

I cannot give users the ability to upload/train their own models, correct?

Disallowed applications: * Help to facilitate payday lending, pseudo-pharmaceuticals, gambling, multi-level marketing, weapons development, warfare, cybercrime, deepfakes, spam, non-consensual surveillance, plagiarism, academic dishonesty, or anything that would violate applicable law.

Here I am only interested in what you consider plagiarism and academic dishonesty

For example, can users upload PDF files (i.e. scientific paper) and have the ability to summarize paragraphs on-demand? The PDF files’ processing will happen on our own servers and Open AI API will only be triggered when the user decides which paragraphs they wish to summarize by clicking a button.

Does that fall in the category of disallowed applications?