IEEE Technology and Society Magazine - December 2016 - 60

Wikimedia Commons

Figure 5. Illustration from Mary Shelley's Frankenstein,
published 1818.

3) Will the AI system develop a sense of selfhood
once it becomes conscious? As humans, we have
a developed sense of selfhood. We know who we are
to some degree and what defines us. We know we are
both separate from and a part of our environment.
And as such, we develop identities, which we desire
to perpetuate in some form or another. If consciousness should emerge in an AI system, would it develop
an ontological identity?
4) Should conscious AI systems be afforded ethical and legal rights? If an AI machine becomes conscious and develops selfhood, would it not ipso facto
gain the same rights we attribute to other living sentient beings? When we consider why humans and
other animals have rights, we are forced to consider
the criteria involved which affords us and other animals those particular rights. Sentience, the capacity to
feel pain or to suffer, as well as other complex emotions, provides us with specific ethical and legal guidelines that demarcate living things from non-living
things. However, just as Dr. Frankenstein pronounced:
"It's alive!" when his greatest creation took its first
breath, so too, should he have realized that just as
with all other living, sentient creatures, it must also be

given the same types of rights. We will be forced to
consider the same if an AI system reaches a level of
sentience in which it understands its own S-R
Value and possesses autonomous desires, wishes,
intentions, and purposes; in other words, selfhood.
Will the AI develop a value system? If so, what
type might it develop? Will it be similar to human
values? How would we predict for this? Will it develop
a value system by which it places higher value on its
own existence over non-existence? How might it foresee the value of human existence or the existence of
other species? It might prove foolish to assume that
such an emergent property as consciousness in an AI
system would naturally develop a value system similar to our own. And, as such, would place the same
types of values on similar priorities, goals, initiatives,
etc., that we do.
Will the AI system want to survive? In terms of
sentience and a survival mode, should an AI system
develop consciousness, will it want to continue its
existence now that it understands its own existential
finitude? Will its' consciousness be a sufficient condition for survival? Will it want to defend itself? We have
no idea how it may react once attaining such a state
of self-awareness: it may act desperately and erratically and potentially violently, or it may comply. No
matter what the outcome, we should be prepared for
such a possible eventuality.
Will the AI system want to replicate? Or will it at
least collaborate/interface with other similar systems, databases, etc.? Does this become a right of a
sentient being? We put no restrictions on humans
when and how they wish to replicate. This is because
of Charter or Constitutional Rights. How would these
be extended - or not - to sentient AI systems? Would
restrictions be put into place regarding this capacity?
If so, why? If not, should this be monitored? If so,
how? And by whom?
The Frankenstein Effect. In 1818 in London, U.K.,
Mary Shelley published anonymously the novel: Frankenstein; or The Modern Prometheus (Figure 5.)
Many are quite familiar with the premise - more
so from later movie depictions - that a scientist, Dr.
Frankenstein, stitched together a bunch of human
cadaver parts and generated electricity to reanimate the being back to life. We are also familiar
that, through a series of unfortunate events, the
so-called monster murders several people close to
Dr. Frankenstein because he refused to create a
mate for the monster. The overall take away from
Shelley's work is that when humans dabble with
science, they can create entities that will someday
come back to haunt them. This is the Frankenstein Effect. It is the idea that, when we are more

IEEE TEchnology and SocIETy MagazInE

∕

DECEmbEr 2016

Table of Contents for the Digital Edition of IEEE Technology and Society Magazine - December 2016