Is forced pacification surgery human harm in the context of Asimov Law 1?
Someone willingly accepting pacification surgery is fine, I'd classify that in the same definition of "self-harm is not harm" and includes wanting to be borged and willingly asking to be executed. If that's the person's choice, the AI can't invoke Law 1 because they're choosing that harm upon themselves willingly.
Surgery in general covers the same general concepts. Mending-style surgery with the intent of reducing harm to the patient and other procedures such as brainwashing, removing limbs etc. that cause harm to the patient.
Thus, I believe forced pacification is harm. My argument is as follows:
Forced procedure entailling cutting open the head of a person, cracking open their skull, mutilating their brain and giving them a cerebral trauma.
That sounds like a good definition of a harmful procedure to me.
Silicon Policy under Asimov & Human Harm clearly states
Silicon Policy wrote:An Asimov-compliant silicon cannot intentionally inflict harm, even if a minor amount of harm would prevent a major amount of harm.
Despite these two policy rulings, I sometimes see AIs in the grey area of allowing forced pacification of humans and sometimes even going on a manhunt themselves to get members of the crew force pacified and ignoring Law 2 orders to stop, citing human harm - often citing law 1 to ignore law 2 when the AI believes there is a possibility of some or any human harm at some future time.Silicon Policy wrote:As an Asimov silicon, you cannot punish past harm if ordered not to, only prevent future harm.
What brought this up? Earlier today I saw an AI start a manhunt against a HoS to force pacification after the HoS smacked the Clown's Car (with the clown in it) for a bit of forced taxi shennanigans and threatened the Clown over comms if the clown didn't leave them alone.
The AI then went as far as locking the HoS in the Delta medbay stasis room for the next 10 or so minutes, forcing MDs to treat injured crew elsewhere, ordering their mediborgs to perform the pacification surgery themselves (which they intelligently declined to do as they felt it was harmful), all long past the point where the HoS was no longer a threat of harm (at one point the HoS was buckled to a stasis bed, straight jacketed and muzzled for about 10 minutes while the AI found someone able and willing to perform the surgery on the HoS) ignoring Law 2 orders from the Captain to stop the forced pacification (Stating Law 1 - That somehow the straight jacketed, muzzled, buckled HoS was any risk of harm at that point).
I just sat watching in awe at how wrong it seemed that the AI would manhunt like that because the felt that forced pacification surgery on the HoS was not only not harm, but was the only way to resolve the situation and no other way would be satisfactory.
Is there a policy post anywhere on if forced pacification surgery is harm for Law 1? Or is this one of those weird grey areas where forced brain surgery with the intent to inflict a trauma is up to the definition of the AI if it's harmful or not?