The AnthroVision Journal special issue on “Computer
Vision” explores design, co-creation, and labour with image recognition
technologies, and the shifting ontologies between knowledge and the senses
using new digital tools. What methodological frameworks are there for
anthropologists to work alongside engineers, designers and other professionals?
We are seeking papers dealing with such issues, as well as, on the conditions
of immaterial labour to create training sets.
Based off "Training Humans" by Dr. Kate
Crawford and Trevor Paglen, the current practices for creating training sets
for computer vision AI harkens back to the colonial era of anthropology:
systems-based interpretations of discrete cultures and the positivistic
apparatus of observational film. In particular, people of color, migrants, and
low-wage workers are the most vulnerable targets of this visual taxonomy.
Furthermore, platforms for training computer vision, such as Amazon Mechanical
Turk, are exploitative. Workers, based mainly in the global south, have just
seconds to analyze each image in order to work at a pace that can profit them.
This complicates the multi-sited entanglements of subjugation and exploitation
between the observer and observed, laying the ground for examining the interrelations
of epistemology, labour and AI bias.
How can anthropologists articulate ethical issues
between knowledge formation, scientific institutions and neoliberalism. How do
anthropologists find reflexive modes of analysis? Where are possibilities for
future interventions?
Send abstract to: jielianglin821@gmail.com and nadinewanono@gmail.com
Abstract Deadline: Monday, September 7th.
Abstract length: 500 words.
Essay length: 6-7000 words
No hay comentarios:
Publicar un comentario