So I created a small project on https://aivshuman.thedaviddias.com/ which parses commits and try to find trace of AI tools. Not 100% accurate, it depends on people commiting using AI.
That's only when you can potentially determine that AI tools were used. It's interesting and triggers some questions about if we should have a better or stricter way to measure AI vs Human produced code.
Feel free to share any thoughts or improvements! The project is open source (https://github.com/thedaviddias/ai-vs-human)