Please describe your role in, or in relation to, software projects that accept or contemplate accepting LLM-generated code.
Do you require any additional documentation, including legal documentation, from contributors about the use of LLMs in generating code? If yes, please describe the documentation.
Do you provide any information or guidelines on how LLMs should or should not be used by contributors?
Are there any specific types of code that are acceptable (less risky) to be generated by LLMs? (configuration/install scripts, datasets, etc.).
Do you require any specific information about LLMs or prompts used from developers? If yes, please describe the requirements.
What additional information do you require in code that includes LLM-generated contributions?
Do you have any safeguards in place if someone claims copyright or license violation on LLM-generated code?
Do you recommend that the LLMs used follow any specific criteria? If yes, please describe the criteria. (e.g., free documentation, specific training data, the LLM's license).
What do you think of the impact of the US Copyright Office
guidelines on accepting LLM-generated source code contribution?
Is there any other issue with LLM-generated code that is relevant from the point of view of your project?
If you can, please share the names or URLs of the projects you are involved with. We will not publish this information.
