There is a project that tracks/curates "AI" in the context of bad use or real world (mis-)application [1].
[1] https://incidentdatabase.ai/