I’m aware that the entire MusicBrainz project is having a lot of problems with AI scrapers, leading to reduced performance and even outages.
Despite this, MusicBrainz Picard seems to be using agentic coding as part of the development process and there is a blog post from an “Agentic-AI FullStack Developer” on the MetaBrainz blog.
For what it’s worth, this is a summary about the student’s GSoC project. For these projects the use of AI is meant to be very limited, and my understanding is that that is the case here (that it was used only to make repetitive changes in files which were then manually reviewed). In their general work outside the project, the students are more than welcome to use whatever tools they want and their future customers are happy with - if that kind of description is what potential employers are hoping to see, then I can see why they’d use it.
We do not have a 100% fixed policy yet for these things, other than what @chaban linked above; the main point is “if AI is used, the use needs to be clearly declared and the developer needs to be in charge and own the changes”. I expect that policy to develop further in the next few months under our new Executive Director. This year we have had a lot bigger fires to fight with the passing of our previous ED, so that short basic thing is all we have arrived on.
At the moment, there are different levels of usage in the team. Some of us don’t use AI at all, some use it for tests writing and benchmarking, and some use it more widely.
Even as a reasonably AI-skeptic person though (I am one of those who has never used AI for code at all), I don’t think it’s reasonable to compare the use of rogue bots stupidly and ineffectively scraping freely available data for AI and causing havoc in the process with a developer who knows what they are doing using some assistance while still owning the results.