In René
Descartes’s Discourse on Method (1637), he
defines the human being as a thinking reality. His Latin statement,
“cogito, ergo sum”, is usually translated into
English as "I think, therefore I am.”
However,
for anyone to expect an A.I. chatbot to say this in the present
indicative (“I think” and “I am”)
and to expect this chatbot to be some insightful budding Hamlet
meditating on what it means “to be” is certain
to be disappointed.
Let’s
be clear, artificial intelligence with its silicone neural-network
is not really intelligence. It’s actually more about
scanning mountains of data and “doing” something
with it (rather more in the light of pattern recognition)
than thinking. And in no way is it the same as really knowing
things, of expressing a living self and the joy of being alive
or knowing that finding a purpose in life is a lot harder
than it looks.
With generative A.I., a new entry in the digital mind-field,
we have created a cognitive amoral machine whose goals determined
by some carbon-based human’s algorithms are said to
echo those fearsome dangers once described in ancient Greek
myths. In particular, four characters, Prometheus, Cassandra,
Pandora, and Narcissus, have once again taken center stage
and are living large in our tech-world.
Prometheus, the first tech entrepreneur, in defiance of the
Olympian gods, stole fire (a metaphor for technology) and
was caught giving it to humanity.
Of course, the gods kept fire from man because they didn’t
trust humans,and considering that Homo sapiens is the only
species that systematically kills its own kind, who could
blame them? Clearly, Prometheus is the mythic stand-in for
the liberating power of knowledge and the threat of technological
overreach and self-destruction. Furthermore, with generative
A.I. as the new fire, this is surely our modern Promethean
moment. Thus, with the world’s cache of nuclear weapons
and accessible software for killing by machine, the question
is how much autonomy do we grant to A.I. in running next-generation
lethal weapons?
Then again, much as we need guardrails and absolute control
of the A.I. product, we should keep in mind the ironic insight
of the Hungarian-born polymath and a pioneer of A.I., John
von Neumann, who said “there is no cure for progress.”
Cassandra, on the other hand, was a Trojan priestess fated
by the god Apollo to speak true prophecies, usually of impending
disasters, but never to be believed.
Our recent Cassandras are the six CEOs from Anthropic, OpenAI,
Google, Microsoft, Amazon, and Meta, who have visited our
president and other world leaders to alert them to the dangers
of A.I. stalking our future. As for the 'doomers,' those true
believers in imminent disaster from A.I., the jury is still
out.
And then there is Pandora, whose name means ‘all-giving,’
and not in a positive way. As a wedding gift, Zeus (king of
the gods) gave her a box (or jar) containing all the evils
of the cosmos. Apparently, angry with Prometheus for having
stolen fire, Zeus decided to punish mortals and to use Pandora
as an instrument of his wrath. Despite being warned never
to open that box, Pandora, curious to a fault, looked within
and unleashed all possible miseries on mankind; yet, strangely
enough, there was one gift left behind – hope. And we
can only.
Hence, the question is whether these generative A.I. algorithms
and the imminent super-intelligent AGI, (Artificial General
Intelligence), will create responses that include some truly
Pandoric ‘out-of-the-box’ existential threats
to humanity. While we are not alone in thinking that, consider
the following ominous comment:
Mitigating
the risk of extinction from A.I. should be a global priority
alongside other societal-scale risks such as pandemics
and nuclear war.
Bill Gates (Cofounder of Microsoft) and Sam Altman (CEO
of OpenAI)
Lastly, there is Narcissus who is forever engaged in a truly
impossible lovers’ quarrel with himself. As a cautionary
tale of self-obsession, he is both unable to resist falling
in love with his own reflection and unable to empathize with
other humans. Thus, Narcissus can be seen as an A.I. surrogate.
In other words, with generative A.I. rewriting its own code
for its own purposes, we may have lost control over our own
algorithms through a silicone “self-centered”
disorder.
Of
course, while we can say we are what we make and that A.I.
is merely an avatar for us, the sense of self-awareness or
intentionality can never be present in A.I. As a result, AI
suffers from an unconscious narcissism and zero empathy. By
comparison to what makes us human, there is nothing less real,
less personal, less poetic or more unpredictable than some
chatbot captive to an algorithm and the “mindless”
rush of electrons. And I should think that is not a good thing.
