[QuadmasterXLII]

not so clear whether this matters for harm intensity- anything an ai can be induced to do by a human, which it can then do at scale, some human will immediately induce. especially if its harmful.