[iDC] MTurk project - introduction

[francesco gagliardi]

Hi everyone,

I’ve been on the list for a while, but I don’t think I ever introduced
myself. I work in performance and occasionally film and video, and write
about performance history. Trebor asked me to introduce to the list the work
I will be presenting at the Digital Labor conference in November. I can’t
say too much about it yet since it is still developing, but here are the
basics.

The piece will be based on Amazon’s Mechanical Turk. I was struck in
learning – through this list I believe – that, according to a recent survey,
a good number of (western) MTurk workers engage with the tasks crowdsourced
through the service in order to kill time and have fun, rather than simply
to earn money. Given the outrageously low wages paid for these tasks, this
should not have come as a surprise. Nonetheless, I was intrigued by the
notion of engaging in MTurk “HITs” (Human Intelligence Tasks) as a form of
entertainment. More specifically, I was intrigued by the performative
quality of this mode of engagement, in which repetitive undemanding
activities meet a pervasive habit of compulsive multitasking.

The original plan was to commission the execution of a series of tasks
highlighting the performative aspect of this type of labor. After looking
closely at a number of projects similarly engaged with Amazon’s service,
however, I began to feel increasingly uncomfortable at the idea of
positioning myself as a task requester, albeit in the context of a critical
project. This discomfort is rooted, I suppose, in a more general skepticism
about the effectiveness of critically confronting a problem by way of
instanciating it, a strategy that nonetheless seems to have a sound standing
in some sectors of the art world, and which occasionally has produced
compelling work (I’m thinking, for instance, of the work of Santiago
Sierra).

I decided that I would develop my project entirely from the standpoint of
the “provider”: I enrolled on MTurk as a worker and have been performing
HITs for the last month or so trying to get a sense of how the platform
works, what kind of tasks do people request, and what is it like to engage
in this kind of work. One of the first things that struck me is the variety
(and oddity) of tasks crowdsourced through the service. Some tasks are quite
straightforward: finding and matching information, translating text,
transcribing audio, cataloguing images, answering surveys. Others verge on
the scam: visiting a blog or a website and living “positive feedback”.
Others are so odd that they sound like they could in fact be artists
projects (“Write about a never before seen living environment”) or cognitive
experiments (“Name all the objects in this photograph”).

I am considering different possible forms for the project’s final
presentation. One way to proceed would be to embrace fully the performative
aspect of this way of engaging with HITs and to devise a way to document
this online work AS performance. Another way would be to appropriate and
subvert material from the HITs (images, instructions, text, lists, etc.) and
use it to produce performance “scores” or “scripts” in the spirit of Fluxus
instruction pieces: short texts, images, or drawings prompting open
performative realizations.

I will be looking forward to meeting some of you at the conference, and to
lively and stimulating discussions.

All the best,

Francesco
-------------- next part --------------
An HTML attachment was scrubbed...
URL: http://mailman.thing.net/pipermail/idc/attachments/20090925/b433a1a7/attachment.htm