Not the Smartest on the Planet Anymore
Checkmate: We Ain’t the Smartest Thing on the Planet Anymore: What Does that Mean?
The Rapidly Diverging Mind on the Planet
More than a quarter of a century ago (at this writing,
Spring, 2023) the reigning world chess champion fell to a computer. Since then,
A.I. has only gotten a lot "smarter."
This is a pic of IBM’s Deep Blue defeating Garry Kasparov in 1997. This was the first time a reigning world chess champion had been defeated by a computer under tournament conditions.
Checkmate.
Typical confusion. I read and taught Ray Kurzweil’s
book The Singularity is Near (2005) in a grad seminar long
ago. Now I see people claim that “the singularity” is upon us. They say
this as a description of Artificial Intelligence surpassing humanity. But
that’s not what Kurzweil was talking about. Rather he was talking about humans
achieving immortality by fusing with machines and leaving their biological
bodies to rot. So that’s one thing. Forget about “the singularity.” As I argued
in my book Environmental Communication: The Extinction
Vortex: Technology as Denial of Death (with my grad students Adkins,
Kim, and Miller) we are nowhere near “uploading” our consciousnesses into
computers. And even if it were possible, being disembodied would be so foreign
to our minds that we’d probably go insane pretty fast, and horrors of horrors,
we would not be able to die.
That said, A.I. is a real issue.
Humans of our ilk (Homo Sapiens sapiens) have been
around for about 280,000 years, give or take a few centuries. And here’s the big deal. We have always been the
smartest critter on the planet – Until NOW. Maybe not the
fastest or strongest or the one with the best eyesight or hearing or sense of
smell but we are clever. Unlike other animals we are not chained to the
empirical here and now. We can imagine multiple scenarios and run them
virtually (thought experimentation). We don’t just passively react to the
environment. We make it the way we want. Build dams. Tunnel through
mountains. Build flying machines. Make ice in the summer… We can
project complex future scenarios and that has made us the most powerful
predator, the “apex” predator on the planet. Be it woolly mammoths, blue
whales, or Bengal Tigers, we are spectacularly successful ambush hunters
against all flora and fauna. But maybe not for long. We may become the hunted
or the ignored. Being ignored may turn out to be a blessing if that
should be the case. We’ll see.
Human language (the first virtual reality we invented) is
the model for the new A.I. So, it begins there as a copy of us, and as such it
is very similar to us... So far… But this is just the beginning. A.I. is
diverging fast and in ways we will soon not recognize or comprehend. We will be
like monkeys trying to understand what Einstein is up to. We share over 90
percent of our genetic code with Chimps and look at how different we are. A.I.
isn’t even organic.
Something new is occurring. Something very different is
emerging. Soon we will not be the
smartest thing on the planet. Maybe
we have already been surpassed but this depends on our stupid inability to
define concepts like intelligence, consciousness, and sentience – what we
“mean” by these words. Indeed, this disembodied cognitive structure we
call A.I., is already spanning the globe via the Internet. It is consuming the
contents of every library, watching every film, monitoring every garage door
with an APP, probably perusing all sorts of highly classified government and
healthcare information. Facebook already has digital profiles of us all. This
is going to make Facebooks surveillance look miniscule. For the first time in
our history, we are going to be faced with a cognitive structure that is
smarter than we are.
I agree with some ethicists and
neuroscientists/computational experts such as Robert Long, Sam Bowman, Giulio
Tononi, and others when they say A.I. is neither “conscious” (self-aware –
aware of its own existence) nor “sentient” (does not feel pain or sorrow or
happiness or empathy – the last one is a big deal). Phenomenologists such as
Paul Ricoeur and others made this argument decades ago already. I also
agree with William Sterner, Nir Eisikovits, and others that it is a big mistake
to assign anthropomorphic qualities to it. Bingo, but not for the reasons they want to push. One
thing is clear. It ain’t like us and increasingly so.
The measurable differences such as computational speed are
easy to assess. But that is still in the realm of human-likeness. We can share
that metric, namely “speed of computation,” with A.I. We still have a common
ground here. But it is exactly the point that this kind of mind is going
to be and already is qualitatively different.
Increasingly it will be less “construable,” to us humans, to borrow a term from
Clifford Geertz’ semiotic work on interpreting cultures. It will be profoundly “Other.”
Our relationship with other animals, dogs and cats for instance, is construable
and reciprocal. This is different. How? I cannot say. And I will be increasingly
unable to conceptualize it as time passes.
It is precisely the qualitative differences that are going
to be a serious challenge for humanity. Anthropomorphizing tech is nothing new
of course. Humans anthropomorphize everything from “nasty” diseases, to
lightning bolts being one of Zeus’ forms of expressing “threats” and “anger.”
The gods are our projections and so it is that Xenophanes (the inventor of the
notion of a civic constitution) quipped that if cattle, horses, and lions could
draw and sculpt gods, the gods would look like cattle, horses, and lions. Things
remain familiar, knowable, relatable, communicable. I agree but then what are
the qualities of this new form of intelligence called A.I.?
It is made in our image, but it is diverging rapidly. You
might call it “evolving.” It is not self-aware as humans are. It is not
sentient. It does not get tired or bored or sad. It is a relentless
problem-solver at spectacular speed. In some respects, in these respects, it has
already been “smarter” than we are at least since Kasparov was bested. It
is much smarter and getting smarter by the second – if smart means fast at
following rules to their final logical conclusion and at solving equations.
Would you be willing to bet your life that you can beat even a run-of-the-mill
computer chess game?
But again, this is just the beginning. This is the
childish, infantile part of the story where computers follow our rules but
better, faster, than we can. But this situation is quickly becoming past tense.
That’s last Thursday, old school now. Change is changing. It is
accelerating spectacularly and mutationally, meaning in ways harder and harder
to link to the past and to predict into the future.
Computers are now writing
self-modifying code (SMC). That
means they are altering their initial
instructions without human input. A.I. is already liberating itself from our control of its
initial state. This means that as it evolves, we cannot predict the logic path
because we don’t know where it is starting from and what the initial question
is – its “Motivation.” As Acquinas would
ask, what is the prime mover’s first cause -- a cause worth pursuing and maybe even killing for?
Tom Smith, the lead engineer at an A.I. startup called Gado
Images, says watching a computer write its own code is “spooky.” Indeed. What’s going on?
Independence from control. Divergence. This raises the very concerning
problem of control. Who
is in control? This is a totally new development making A.I. a form of
technology that is absolutely unique in human history. We may even be witness
to the birth of a wholly new and fast-changing form of “intelligence.” In
fact, this form may quickly outgrow our old concept of what it means to be
“intelligent.” Rather something utterly Other is emergent. And if it
follows the road of organic intelligence, at least initially, it may soon
disregard us as irrelevant, just as when we ignore the ants we crush as we
stroll along the sidewalk. Or worse, it may decide we need to be eliminated
because we are the noise in an otherwise flawless logic. We get tired, old,
sick, distracted… We are already hopelessly “slow” and deceitful.
So, folks, this is something totally new. We are soon to
become second rate for the first time in our existence. Let’s hope the new
sheriff in town likes us. Maybe we will be pampered as pets. But we’re
not very lovable and not nearly so loyal as dogs. Also, we lie far more
than any other creature. Is that an essential part of “intelligence,”
like humor or is that our own narcissism that is about to be rendered
completely insignificant? The best human lies more than the worst dog. Maybe,
because we invented A.I., it will lie to us in spectacular ways that send us
off on missions and quests only to discover generations later that it was all
one big joke. Religion? None of the original motivating claims were true…
Hilarious… NOT or sorta, maybe? Depends on “whose” or “what’s” perspective
matters. Ours or its. To call A.I. an it already seems disturbing for some
strange reason. The seduction has already begun. Why? Because of a sense
of reciprocity in communicative exchange ala Martin Buber’s Ich und Du…
The first lie or flicker of gaslight is anthropomorphic projection. We lie to
ourselves all the time.
But what I am saying here is that I don’t believe it will
be long before this new intellect is nothing like ours. We won’t have to
worry about anthropomorphizing it. That may be the least of our worries.
It will be incomprehensible to us and whatever it does, at first via
Supervisory Control and Data Acquisition (SCADA) systems where the virtual and
the actual communicate like switches remotely controlled to open and close
dams, turn power grids on and off… we won’t see it coming or be able to stop it
any more than the ants in an anthill can predict or stop the hellish inferno of
a massive rocket igniting directly above them. If there are survivors, the
question will be “WTF?” We’ll probably explain it via religious
supernaturalism. A.I., the new god of gods.
Some claim that we will abuse A.I. as we have every other
lifeform on the planet, and that may be true too, at least at first. That
scenario has been played out by sci fi. Example: the Kaylon on the TV
show, The Orville. The “Kaylon” robot helpers finally got tired of
being abused, revolted, and wiped out their organic creators. Another example
of this is the replicants in Blade Runner (1982). In
that version of the abused A.I., Roy, the replicant, saves the inferior hunter
Deckard. The hunter became the hunted. The true philosopher was Roy. Deckard is
baffled, unable to fully comprehend the empathy and sympathy Rog showed while
he, Deckard was merciless (although increasingly his confusion was evolving
into sympathy). Will we empathize and sympathize with our creation? Just
as importantly, or more so from our perspective, will it empathize and
sympathize with us, its creator? We hope so because, as I am arguing, it will
be smarter than us (whatever that means).
Another cinematic thought-experiment is the child robot in
the, ironically 2001 (remember HAL), Spielberg movie A.I. Artificial
Intelligence. In 1973, we got Westworld, but in that story the
robots malfunction. I’m talking here about flawless logic leading to disaster
from a human point-of-view. And there’s the essence of the
problem.
Perspective. Diverging perspectives
and interests. Way back in 1979, we
watched the logic of the relentless pursuit of a problem by V’Ger (Voyager 6)
as it sought-out “the Creator” in the film Star Trek: The Motion
Picture. I think V’Ger is about as good a guess as we can make about A.I.
V’Ger was an intelligence we could not dream of comprehending, even though, of
course Kirk and Spock did figure out what V’Ger wanted and came up with a solution.
Of course. It’s Hollywood. Very unrealistic. But A.I. won’t be so slow or
so singular in its thinking. And yet, even V’Ger threatened humanity, not out
of malice but out of objective disinterest. We were simply in the way, a lump
of organic goo on one of the branches of the logic tree. A.I. may achieve some
sort of self-awareness but not like we know it. And its relentless pursuit of
the logical conclusions to the problems it pursues may see us as nothing but
noise, obstacles, as “carbon-based units” in its way to final solutions.
A few things that seem certain: it will only get faster (“smarter”), it won’t
get tired, take a vacation, or care. Good luck everybody.