Humanist Discussion Group, Vol. 36, No. 487.
Department of Digital Humanities, University of Cologne
Hosted by DH-Cologne
www.dhhumanist.org
Submit to: humanist@dhhumanist.org
Date: 2023-03-27 18:07:01+00:00
From: James Rovira <jamesrovira@gmail.com>
Subject: Re: [Humanist] 36.483: followup: agency & intelligence
Thanks very much for this link -- I very much enjoyed reading Krista
Kennedy's article. There are different questions at play, though. I've been
thinking of intentionality and agency in terms of -interpretation-. Not
just to whom or to what do we ascribe the text, but how do our
interpretations of the text shed light on our view of the author? Kennedy's
article doesn't address that question.
I'm going to think out loud about the article as I read it.
I'm not sure that I buy her claims about the authorship of encyclopedia
entries. What she describes is just synthesis writing. It's a review of the
literature in linear research papers. Encyclopedia writing is just an
abbreviated, general form of that task. I like her description of it, but I
don't think she particularly makes a valid point from it. Speeding up the
process in online encyclopedias like Wikipedia doesn't change the nature of
the authorship, except that it's directly collaborative rather than a
single author drawing from multiple sources, but then printed encyclopedia
entries could be coauthored as well.
When she gets into the use of bots to create wiki entries, I don't know
that we're really on new territory there either, as the process she
describes is the assemblage of US census data that was indeed written by
humans. There is no "bot-authored" text in that case. It is all human
authored. It's not even "bot-arranged" by her own description, as a human
being programmed the bot to arrange the text in a predetermined format. I
don't think we're facing any particular complications there either.
I believe she potentially has more of a point with bot editing, but I need
more details about how the bot editing works. If the bots are programmed to
replace "faulty" human authored text with approved text from other sources
(say, US census data), then the bot is replacing humanly-generated text
with other humanly-generated text. Again, no complication here. But it may
be doing more than that. I don't know.
I very much like her distinction between agent and agency: we can't have an
agent without agency, but we can have agency without an agent. But I think
this question could be answered in the case of bots with more detailed
description of the kinds of edits that wiki bots make and how they are
programmed to make them. What she describes sounds like the bot is an
automatically operating tool of a human agent following a human logic, so
that agency resides in the human programmer, who is the agent, and the bot
is then just a tool. My washing machine automatically washes my clothes in
a way similar to how I would wash them by hand, but I give credit to the
inventor of the machine, not the machine itself.
I very much like her legal discussion, but I think her conclusions are
fundamentally wrong. The legal dispute is not over authorship but over
-ownership- of the material under the name of -authorship-. This is a
question of -payment- and -royalties-. There's no question that the
consultant was a writer, author, contributor, etc. of a significant amount
of material in the script. The legal question is only about his credit as a
coauthor of the script. I think his claim is reasonable and fair, but he
didn't enter into that kind of agreement ahead of time, so the ruling was
against him. There's no separation of agent and agency here, just a
separation of writing and ownership. Kennedy is mistaken to equate the two.
I think her description of work-for-hire issues and their extension to bots
is similarly flawed. She seems to be taking a superficial similarity and
making an identity out of it. Work for hire is a prearranged agreement for
a commercially available product. Student writing is not that.
No intent to apply this idea to the article, but I would love every article
about ChatGPT and similar technologies to start out with the following
sentence when it is in fact true:
DISCLAIMER: I really have no clue how this technology actually works.
I wonder how that would change our discourse and discursive environment
about this subject?
Jim R
On Mon, Mar 27, 2023 at 2:52 AM Humanist <humanist@dhhumanist.org> wrote:
>
>
> For anyone following this debate about agency, intention and authorship, I
> can
> recommend Krista Kennedy’s reflections on bot-authored texts in Wikipedia.
> Her
> treatment of the topic is admirably rational and concise:
>
> Kennedy, Krista. “Textual Machinery: Authorial Agency and Bot-Written
> Texts in
> Wikipedia.” In The Responsibilities of Rhetoric, edited by Michelle Smith
> and
> Barbara Warnick, 303–9. Long Grove, Illinois: Waveland, 2010.
> https://surface.syr.edu/wp/1.
>
>
--
Dr. James Rovira <http://www.jamesrovira.com/>
- *David Bowie and Romanticism
<https://jamesrovira.com/2022/09/02/david-bowie-and-romanticism/>*,
Palgrave Macmillan, 2022
- *Women in Rock, Women in Romanticism
<https://www.routledge.com/Women-in-Rock-Women-in-Romanticism-The-
Emancipation-of-Female-Will/Rovira/p/book/9781032069845>*,
Routledge, 2023
_______________________________________________
Unsubscribe at: http://dhhumanist.org/Restricted
List posts to: humanist@dhhumanist.org
List info and archives at at: http://dhhumanist.org
Listmember interface at: http://dhhumanist.org/Restricted/
Subscribe at: http://dhhumanist.org/membership_form.php
