Honestly, I can understand the RPCS3 team's frustration. Emulator development is already extremely complex and low-level, so blindly submitting AI-generated patches without proper testing probably creates more work than help.
AI can still be useful as an assistant for debugging, documentation, or boilerplate ideas, but contributors should at least understand the code they submit and verify that it actually works. Otherwise maintainers just end up reviewing broken PRs all day instead of improving the emulator.