Page 344 - Data Science Algorithms in a Week

Explorations of the ‘Transhuman’ Dimension of Artificial Intelligence 325

For Gelernter, therefore, contemporary AI-research or “computationalism”, disregarding
the other (inalienable) mental focus-levels that humans are privy to, is preoccupied with
rational thought, or “intelligence”, precisely, which is why they believe “that minds relate to
brains as software relates to computers”. (Gelernter 2016: xviii-xix). He compares current
research on the mind to dozens of archaeological teams working on the site of a newly
discovered ancient temple, describing, measuring and photographing every part of it as part of
a process that, they believe, will eventually result in a conclusive report embodying the ‘truth’
about its properties. He disagrees with such an approach, however (Gelernter 2016: 1):

But this is all wrong. The mind changes constantly on a regular, predictable basis.
You can’t even see its developing shape unless you look down from far overhead. You
must know, to start, the overall shape of what you deal with in space and time, its
architecture and its patters of change. The important features all change together. The
role of emotion in thought, our use of memory, the nature of understanding, the quality of
consciousness – all change continuously throughout the day, as we sweep down a
spectrum that is crucial to nearly everything about the mind and thought and
consciousness.

It is this “spectrum”, in terms of which Gelernter interprets the human mind, that
constitutes the unassailable rock against which the reductive efforts on the part of
“computationalists”, to map the mind exhaustively at only one of the levels comprising its
overall “spectrum”, shatter. This is particularly the case because of their hopelessly
inadequate attempt to grasp the relationship between the mind and the brain on the basis of
the relation between software and hardware in computers.
In an essay on the significance of Gelernter’s work, David Von Drehle (2016: 35-39)
places it in the context of largely optimistic contemporary AI-research, pointing out that
Google’s Ray Kurzweil as well as Sam Altman (president of Startup Incubator Y
Combinator), believe that the future development of AI can only benefit humankind. One
should not overlook the fact, however, Von Drehle reminds one, that there are prominent
figures at the other end of the spectrum, such as physicist Stephen Hawking and engineer-
entrepreneur Elon Musk, who believe that AI poses the “biggest existential threat” to humans.
Gelernter – a stubbornly independent thinker, like a true philosopher (he has published on
computer science, popular culture, religion, psychology and history, and he is a productive
artist) – fits into neither of these categories. It is not difficult to grasp Hawking and Musk’s
techno-pessimism, however, if Gelernter’s assessment of AI as the development of precisely
those aspects of the mind-spectrum that exclude affective states is kept in mind – what reason
does one have to believe that coldly ‘rational’, calculative AI would have compassion for
human beings? Reminiscent of Merleau-Ponty, the philosopher of embodied perception,
Gelernter insists that one cannot (and should not) avoid the problem of accounting for the
human body when conceiving of artificial intelligence, as computer scientists have tended to
do since 1950, when Alan Turing deliberately “pushed it to one side” (Von Drehle 2016: 36)
because it was just too “daunting”. For Gelernter, accounting for the human body means