None. Remember that the response is AI generated. It’s probabilistically created from people’s writings. There are strong relations between that command and other ‘dangerous commands.’ Writings about 'dangerous commands ’ oft contain something about how they should ‘only be run by someone who knows what they are doing’ so the response does too.
As it does in many other areas, the controversial nature of the discussion poisons the well on sources supporting either view. The days of ‘here’s a study saying…’ being a useful tactic in anything are kind of dead. Most discussions can have reliable-sounding sources to support contradictory points. It gets hard to find the truth about anything without engaging in in-depth meta-analysis, let alone in a place like a comments section under a webcomic.