[FRIAM] genai and critical thinking

[steve smith]

On 2/11/25 7:43 AM, glen wrote:
> The Impact of Generative AI on Critical Thinking: Self-Reported 
> Reductions in Cognitive Effort and Confidence Effects From a Survey of 
> Knowledge Workers
> https://www.microsoft.com/en-us/research/uploads/prod/2025/01/lee_2025_ai_critical_thinking_survey.pdf 
>
>
> It really doesn't seem that different to me from numerical analysis. 
> It shifts the work from doing the computing to declaring what the 
> computing should do.
>
Spatial Knowledge:

    I guess it's a cheap comparison but I've watched the impact of
    handheld mapping and route planning devices (i.e. consumer GPS) and
    can't say it has actually significantly harmed/changed spatial
    awareness...  Map wonks seem to use dynamic GPS-based mapping apps
    to enhance their existing experience with paper maps printed years
    or even decades ago?   Those with little spatial/map
    awareness/interest seem to have leaned more heavily on the tech and
    possible increased their dependency on the tech without enhancing
    their spatial awareness/interest particularly.

    I know plenty of people who have *always* wanted/needed to be
    topological not geometric... "just tell me how to get there" and
    today they turn on "driving directions" with voice and "turn left,
    turn right, etc." as they are told and as long as it works are at
    least as happy as they used to be when they had a copilot who was
    doing the navigation and happier than they were when they had to try
    to read the instructions scribbled on their napkin as they drove,
    trying to ignore the scribbled (and rarely-to-scale-map-diagram
    drawn ideosynchratically by someone who *already* knows how to get
    there? ).

    If used with a little thought, both modes are enhanced (IMO) by the
    digital/automated support tools.   Although very few automobiles
    have any paper maps in their glove boxes today... so in that sense,
    something was harmed/lost?

Pre-Digital Information Retrieval:

    As we know from my incessant anecdote about "keeping company" with
    LLMs (primarily GPT), I find the conversational mode of "search"
    very compelling.  But I already reformulated my "search" as an
    anticipatory exercise... first with dictionaries and encyclopedias
    and libraries by, for example, pulling several books from the stacks
    at once, ordering them in some mode which helps me work my way
    through the *imagined* or *foreshadowed* path I am anticipating when
    I pull the books.   I usually had a few words queued in my head when
    I first opened a dictionary or encyclopedia which became a
    growing/shrinking queue as I addressed each word on the stack/queue
    in my head.

Conversational Mode of Information Retrieval/Knowledge Development:

    I don't know if I'm half as capable as Glen in developing a good
    "strawman argument" but I do find that the conversational mode of
    constructing my pursuit of information/knowledge/wisdom leads me
    toward something like that...  deliberately asking GPT questions
    which are either minimal in my assumptions or acutely specific about
    my implications.   What GPT offers which *most*
    conversants/correspondents do not is an acutely careful ear to
    listen to my explicit assumptions and an ?unbiased? ability to take
    those assumptions and my questions and convolve them with their
    *very elaborate* knowledge base (training) and give me what are very
    convincing answers.   The challenge then becomes recognizing
    confirmation bias.   It *is* rather easy to ask a question in a way
    which telegraphs to the over-eager LLM to give me answers which are
    tailored to satisfy the sentiment implied in my question(s).

Like all new tools, I do think we are in a phase of accommodation and 
adoption, like the early days of TV when all performers and producers 
could generate for some time was Radio Programs with moving images.   Or 
Cinema as recorded Theater.   Or Hypermedia as magazines and newspapers 
with hyperlinks (and <blink> tags?).

I have used GPT very sparingly in "voice-conversational" mode and have 
been fairly impressed with how well it works.  It is patently not the 
same as chatting with a friend or colleague, nor is it like using voice 
I/O for the written/typed mode.   As those of you who know me in-person, 
my "conversational affect" in person is somewhat different than my 
affect here (for better and worse).

The Internet made me a *more capable* but absolutely less-frequent user 
of paper libraries.   Mobile devices with mapping/GPS has made me a 
*more capable" spatial awareness thinker/navigator/map-reader but 
actually open many fewer paper maps.  In "idiom-conflation", I even find 
myself occasionally wanting to pinch-zoom a map or look for the little 
blue dot to find my "current" location.  Or wish paper journalism had 
hyperlinks in them to make it easy for me to tangent to the 
implications-of or supporting-info for a given line of discussion.

mumble,

  - Steve




-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://redfish.com/pipermail/friam_redfish.com/attachments/20250211/1a1fd57c/attachment.html>
-------------- next part --------------
A non-text attachment was scrubbed...
Name: OpenPGP_0xD5BAF94F88AFFA63.asc
Type: application/pgp-keys
Size: 3118 bytes
Desc: OpenPGP public key
URL: <http://redfish.com/pipermail/friam_redfish.com/attachments/20250211/1a1fd57c/attachment.bin>
-------------- next part --------------
A non-text attachment was scrubbed...
Name: OpenPGP_signature.asc
Type: application/pgp-signature
Size: 840 bytes
Desc: OpenPGP digital signature
URL: <http://redfish.com/pipermail/friam_redfish.com/attachments/20250211/1a1fd57c/attachment.sig>