I disagree with CDM’s call for more explicit definitions in Bandcamp’s new anti-AI policy announcement.  IMHO, Bandcamp have done the right thing by allowing for human judgement in determining what constitutes a violation, and what consequences should follow, for any particular case.  Encoding more rigid definitions for both violations and consequences is an invitation to bad actors to find and exploit legalistic loopholes.  I prefer the broad statement of purpose the announcement makes, coupled with nuanced human evaluation of specific cases.