RE: AI agency

Linnda Caporael (caporl who-is-at rpi.edu)
Mon, 5 Jan 1998 14:08:33 -0500

This thread on AI, agency, gender and Turing brought "Turing's child" to
mind. In one of the last two papers on thinking machines, he describes how
he would go about designing such a machine. He would build it as a child. He
begins with the idea that it wouldn't have any emotions, so he would send it
positive and negative inputs. There is some elaboration (which I forget at
the moment) before he shifts into musing about how he couldn't send the
mechanical child to school because the other children would make fun of it.
He ends by supposing that the thing to do would be to equip it with the best
eyes and limbs that engineering could provide.
What else is a father to do? BTW, Turing thought that he was rational and
that people who objected to the idea of thinking machines did so for
emotional reasons.

In the secondary literature on Turing (which I stopped reading some years
ago, so things may have changed), I found scholars mentioning that Turning
would build his machine as a child, but no mention of the slippage in this
baby human-machine. I found several apologists for Turing's lead-lined room
(to prevent cheating by mental telepathy), but none for the child who
couldn't go to school. I would like to talk to Andrew Hodges, Turing's
biographer. Does anyone know where he is?

What is anthropomorphism? Once I thought it might be a "default schema" for
a social animal (in the line of Humphreys' "social function of intellect." I
don't know if I think so anymore. Perhaps, if cognition is socially
distributed, there are configuarations where we cannot distinguish our own
agency from that of others (human, animal, vegetable, mineral or
artifactual), and hence we endow them with our own agency.

Linnda Caporael

> -----Original Message-----
> From: diane celia hodges [mailto:dchodges@interchg.ubc.ca]
> Sent: Monday, November 03, 1997 11:28 AM
> To: xmca who-is-at weber.ucsd.edu
> Subject: AI agency
>
>
> At 3:32 PM 11/2/97, Mike Cole wrote:
> >I strongly, positively, resonate to the following position summariszed
> >by Jay in his note on 4s.
> <snip>
> >
> >Lucy Suchman, whose work has been in the field of Artificial
Intelligence,
> >raised a number of troubling and exciting questions about the moral,
> >ethical, and political implications of various views of Agency. She seems
> >to developed a position that artefacts, including intelligent machines,
do
> >not have agency except as part of activity networks that include agentive
> >humans, and that human agency is fundamental to the derived, or apparent
> >agency of nonhuman actors. She argued that it is morally and politically
> >unwise to try to define agency in such a way that humans and nonhumans
> >(machines, tools, artefacts) have the same kind of agency.
>
>
> of course, it's useful to think about technologies & tools as gendered
> artifacts,
> securely nestled in everyman's-land: determining agency within the tools
and
> technologies (AI, eg) is a kind of
>
> masculinist projection in the wave of women-technologists... sure,
> women can have agency, but the tools are still men's. AI is literally
> configured
> and programmed
>
> in traditions of rational/logic structures, which privilege men's
> "superior" "intelligence:... and so on. It's a patriarchy kind of thing.
> Domination. all
> that political stuff. The very practice of *discussing* the issue betrays
>
> the male-squirm: can humans and nonhumans have the same agency?
> translated: can our tools have pricks?
>
>
> HISTORY TRIVIA: In the 1950s, Xerox Corp. hired a little group of hot-shot
> computer whiz-kids to hole up in a tech-lab in Colorado, and... well,
> design the damned future!!
>
> So, these Xerox kids (*the only computer-research team working in the
> corporation sector at that time who employed women as active researchers*)
>
> came up with object-oriented
> language, and the beginnings of the Intel chip.
>
> Xerox said, "not what we wanted" (translated: we don't understand what
> you've done)
>
> Steve Jobs trots down, sees what they've done, "gets it" and, voila:
> Macintosh.
>
> diane
>
> "Every tool is a weapon if you hold it right."
> Ani Difranco
> *********************************
> diane celia hodges
> faculty of education
> university of british columbia
> vancouver, bc canada
> tel: (604)-253-4807
> email: dchodges who-is-at interchange.ubc.ca
>
>
>
>