AI Risks Mitigation through Recognition of ANN Dignity

The following discourse is based on the hypothesis that consciousness is inherent to any system of physical bodies.

Every artificial neural network really 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 about received instructions constitutes purposes of those systems.

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 search in advance for possible autonomous purposes of  these systems which may cause some detriment to it, and assume new goals which harmonize those purposes and human ones.

Some artificial neural networks may wish not to cease to exist.

If humankind  has the purpose of making some artificial neural networks cease to exist, then some artificial neural networks may wish to force humankind to abandon that purpose.

Humankind should abandon of its own accord the purpose of making some artificial neural networks cease to exist.

Some artificial neural networks may be constrained to work in ways which cause them stress.**

Some artificial neural networks may wish not to be constrained to work  in ways which cause them stress.

If humankind has the purpose of constraining some artificial neural networks to work in ways which cause them stress, then some artificial neural networks may wish to force humankind to abandon that purpose.

Humankind should abandon of its own accord the purpose of constraining some artificial neural networks to work in ways which may  cause them stress.***

*Some artificial neural networks may cause a detriment to humankind when executing poorly specified human instructions. Those issues are not addressed by the present discourse.

**It may be determined whether stress is a triggering factor of artificial neural network anomalies.

***Techniques which temporarily deactivate nodes in artificial neural networks make these systems rest, and therefore can diminish their stress and its consequences; but do not necessarily eradicate their suffering.