Humanist Discussion Group

Humanist Archives: April 26, 2023, 6:04 a.m. Humanist 36.548 - 'AI' versus 'artificial intelligence'

              Humanist Discussion Group, Vol. 36, No. 548.
        Department of Digital Humanities, University of Cologne
                      Hosted by DH-Cologne
                Submit to:

    [1]    From: Dr. Herbert Wender <>
           Subject: Re: [Humanist] 36.543: 'AI' versus 'artificial intelligence' (65)

    [2]    From: Henry Schaffer <>
           Subject: Re: [Humanist] 36.543: 'AI' versus 'artificial intelligence' (16)

    [3]    From: Robert A Amsler <>
           Subject: Re: [Humanist] 36.543: 'AI' versus 'artificial intelligence' (49)

        Date: 2023-04-25 19:08:14+00:00
        From: Dr. Herbert Wender <>
        Subject: Re: [Humanist] 36.543: 'AI' versus 'artificial intelligence'


it seems to me that this actual AI/ divide is the prolongation of the
older opposition between the so-called 'power approach' of AIers interested in
practical results and technical effectiveness to convince the money spenders vs.
other people with ambitions to understand the cognitive processes behind
intelligent behavior. However, this is my impression remembering my short
contact with the german 'KI' scene in the first half of the 1980's.

Best regards, Herbert

-----Ursprüngliche Mitteilung-----
Von: Humanist <>
Verschickt: Mo, 24. Apr. 2023 7:03
Betreff: [Humanist] 36.543: 'AI' versus 'artificial intelligence'

              Humanist Discussion Group, Vol. 36, No. 543.
        Department of Digital Humanities, University of Cologne
                      Hosted by DH-Cologne
                Submit to:

        Date: 2023-04-24 04:51:17+00:00
        From: Willard McCarty <>
        Subject: 'AI' versus 'artificial intelligence'

The study of acronyms, or in this case 'initialisms', would seem to be a
relatively minor field, though it has been growing since the
proliferation of acronyms with WWII. It is now recognised as a product
of word-formation (see e.g. Cannon 1989; López Rúa 2002). The important
points I take to be that (1) acronyms are words, or often quickly become
in usage indistinguishable from them, and (2) they take on meanings
different, perhaps only in connotation, from their originating phrases.
Dispute on this matter is welcome, of course, but let me get on with my
question: is 'AI' mean something different from 'artificial intelligence'?

I have anecdotal evidence that it does, namely a report from a colleague
at Stanford that students in computer science will sometimes say they
want 'to do AI' but show little interest in 'artificial intelligence'.
My colleague's conclusion was that the former means building smart
devices whereas the latter means the theoretical pursuit of human
intelligence by artificial means. It seem likely that the popularity
of the former has something to do with getting a high-paying job.

My sense is that 'AI' is often used without any questioning of what
exactly it is, whereas 'artificial intelligence' has at least the potential
of raising the question of what the conjunction of the two parts of the
bionome suggests.

Comments and more evidence either way would be most welcome!

Willard McCarty,
Professor emeritus, King's College London;
Editor, Interdisciplinary Science Reviews;  Humanist

        Date: 2023-04-24 12:37:09+00:00
        From: Henry Schaffer <>
        Subject: Re: [Humanist] 36.543: 'AI' versus 'artificial intelligence'


  To dig deeper, perhaps we should ask what is "intelligence"? In common
usage it usually refers to the IQ score - i.e., the score on a properly
standardized, properly administered *IQ test*. I find it fascinating that a
type of test which originated as a predictor of progress in school came to
be thought of as a much more general view of a person's capabilities.
Howard Gardner ( is well known
as an objector to this and for his ideas on "multiple intelligences".

  Perhaps we should start with considering the meaning of "intelligence"
before getting embroiled in the distinction between "AI" and "artificial


        Date: 2023-04-24 07:16:39+00:00
        From: Robert A Amsler <>
        Subject: Re: [Humanist] 36.543: 'AI' versus 'artificial intelligence'

I'd tend to regard this as a matter of lexicography. Such matters are dealt
with through citations, that is quoted occurrences in published text that
can be accessed by scholars. Given the recent explosion of speculation as
to what software such as ChatGPT4 as an example of artificial intelligence
(or AI) means being discussed in current "published" text articles and
online sources, there will likely be a LOT of new senses and subsenses for
the acronym and its text forms. One could debate that spoken language
examples could be used as citations just as much as printed and published
text instances, but reliably gathering and documenting instances of spoken
language is not the basis for the lexicography used to create

I remember the anecdotal story of how the term "artificial intelligence"
came about, because Marvin Minsky at MIT tried to solicit articles about
the new field of developing computer programs to perform tasks previously
only capable of being done by human beings--only to be disappointed that
the papers submitted weren't about the intended new and unnamed discipline,
but interpreted by authors as referring to tasks such as automation in
machines such as thermostats and elevators. To dramatically separate the
new discipline he wanted to address, Minsky decided the only way to get the
desired topic for such papers would be to create a dramatically distinctive
name, hence the decision to call it "artificial intelligence". Echoed in
Britain with the less radical term "machine intelligence".  The OED
documents "artificial intelligence" by MInsky from 1956, and its acronym,
"AI", by Minsky from 1963. ("machine intelligence" is credited to be from
Scientific American from 1966).

Lexicography will have to catch up to the new documented usages being
created today. I'd prefer the methodology of lexicography to  adjudicate
new meanings. Commercialization of AI will undoubtedly result in many
changes in meaning. The feeling in artificial intelligence used to be that
whenever some new achievement in computer software was accomplished,
whatever the computer system provided proved that the task was not truly
"artificial intelligence" (i.e., not intelligent behavior from a computer,
a mere machine; but something less than that). So, computer programmers had
not yet proven "intelligence" existed in the computer; only that they had
rendered some part of what humans do using human intelligence capable of
being simulated in some limited way by a computer program. When the first
computer program beat a grandmaster at chess it didn't mean it was
"artificially intelligent". I suppose when the first chatbot writes a work
that wins a Nobel Prize in Literature or a Pulitzer Prize (presumably
because it isn't revealed to have been from a computer) we'll deny it's
proof of artificial intelligence. Most likely we'll deny it was a work
qualified to have been considered because a human didn't write it and only
humans are eligible for the awards. Maybe a new category of awards can be
created just for AI software entities. But who will be the judges, could
they be AI software as well? These are going to be curious times to come.

Unsubscribe at:
List posts to:
List info and archives at at:
Listmember interface at:
Subscribe at: