guidelines on using generative AI in contributions?

[gernot-h]

I just stumbled over these nice suggestions in an SW360 GSoC application: eclipse-sw360/sw360#3631 > "Usage of AI in coding".

I think it would be worth to include similar guidelines to https://github.com/siemens/debsbom/blob/main/CONTRIBUTING.md.

[Urist-McGit]

I do not think we need a special guideline for AI contributions. The projects that introduce them are usually dealing with too many AI-generated low effort contributions, overwhelming the reviewers. This is definitely not a problem we have.

The only expectation I have is that the contributor themselves is human and available for discussion. Maybe we should document that, but I think that is an implicit assumption in most Open Source projects.

[gernot-h]

Well, I also thought our small https://github.com/sw360/capycli wouldn't run into this topic, but we did. However, we were able to handle it (after patiently trying to explain our feedback over and over) basically by asking the AI operators to
a) describing their local test setup and how they tested their changes and b) providing test cases for their changes, assuring all tests stay green.