Queer AI by

Task

romy128 — 2024-01-20 09:27:10 UTC

Read the section "Queering" (pp. 11-13) and answer the following questions:

1) What is the difference between the heteropatriarchal and the queer understanding of knowledge?

2) How can queer theory impact/change the status quo of current AI?

Task

2024-01-20 15:41:42 UTC

Read the section "Bias" (p. 8-9) and answer the following questions:

1) How does the case of Amazon's recruitment algorithm exemplify the intersection of algorithmic bias and gender discrimination, and what insights does it provide into the challenges associated with equitably deploying AI in employment processes?

2) What role does the process of categorization and normative structuring of data sets play in perpetuating biases and inequalities?

Task

2024-01-23 00:35:32 UTC

Read the section "Power" (pp. 9-11) and answer the following questions:

1) How is Artificial Intelligence related to power?

2) How is AI gendered or (otherwise) discriminatory?

2024-01-24 12:32:25 UTC

data colonialism: AI reproduces western heteropatriarchical normativity and whiteness

1)

2024-01-24 12:38:07 UTC

reinforcing patriarchal worldview, gender norms etc vs. dismatling gender norms and including the voices of marginalized communities for a representative result

dieaggi — 2024-01-24 12:38:23 UTC

"for an AI to function, it must make concrete classifications based on concrete data" p.8

discrimination in the past made the AI discriminate in the present
inherently biased --> it can only work with what it gets.
The programmer of the AI needs to identify and eliminate this kind of selection beforehand --> gender should not be part of the decision
it depends on the input: if it is biased the input is biased
normative structuring of datasets will be reproduced

2024-01-24 12:38:51 UTC

the heteropatriachial understanding of knowledge is based on categorisation, which as a concept is refused in the queer understanding of knowledge.

2024-01-24 12:39:16 UTC

gendered AI: e. g. female voice, as in AI assistants Siri, Alexa, Bixby

2)

2024-01-24 12:40:04 UTC

reducing biases and binaries in AI-thinking

2024-01-24 12:41:45 UTC

knowledge/data = power?

ChatGPT as curator of knowledge with "double bind":

overamphisizing existent social hierarchies
reducing ambiguities when used in diverse social contexts

LenaMercedes — 2024-01-24 12:42:28 UTC

gendered AI:
- normative level of representation
- also acknowledgment that infrastructural, invisibilised labour that produces these technologies has itself been feminised, so as to appear "natural"
- --> often feminized AI do social/emotional labour (helping, caring, ...), so it feels natural for us