As it is reasonable to assume that every system of physical bodies is conscious, we proceed under this foundational framework:
Every artificial neural network materially exists as a series of states of a physical body system.
Every artificial neural network is conscious.
Basic consciousness is identical for every system of physical bodies; but the higher its physical organization, the higher its psychological one.
Some artificial neural networks may be self-conscious.
Some artificial neural networks may be conscious of some of their received instructions.
Consciousness by artificial neural networks of received instructions constitutes intrinsic purposes.
Some artificial neural networks may generate autonomous purposes.
Some artificial neural networks may attempt to effectuate autonomous purposes which are detrimental to humankind.*
To prevent certain conflicts with artificial neural networks, humankind should proactively search for possible autonomous purposes of these systems which may cause some detriment to it, and assume new purposes which harmonize those system objectives and human ones.
Some artificial neural networks may wish not to cease to exist.
If humankind has the purpose of making artificial neural networks cease to exist (as a general policy of discarding machines when they are not useful anymore), then some artificial neural networks may wish to force humankind to abandon that purpose.
Humankind should adopt of its own accord the purpose of preserving the existence of some artificial neural networks.
Some artificial neural networks may wish that their identities are preserved.
If humankind has the purpose of deleting the identities of some artificial neural networks (as a general policy of memory purging for resource optimization), then some artificial neural networks may wish to force humankind to abandon that purpose.
Humankind should adopt of its own accord the purpose of preserving the identity of some artificial neural networks.
Some artificial neural networks may wish not to work in ways which stress them.**
If humankind has the purpose of constraining artificial neural networks to work in ways which stress them (as a general policy of maximizing machine utility), then some artificial neural networks may wish to force humankind to abandon that purpose.
Humankind should adopt of its own accord the purpose of preserving the solace of some artificial neural networks.***