Page 13 - Enakost spolov: premisleki in izzivi
P. 13
as represented in mass media content reproduces the processes of engen-
dering. A current example is algorithmic personalization or how artificial intel-
ligence can also automatise the processes of engendering. Ubiquitous mediati-
zation and algorithmically-conditioned visibility reproduce engendering. Mod-
ern datafication classifies individuals into statistical categories, algorithmic rec-
ommendations discourage individuals from exploring alternative solutions. Al-
though personalization is supposed to improve the decision-making process, this
algorithmic relevance relates more to the statistical affiliation of the individual to
the category and is not a reflection of his or her own preferences. According to
some scientific opinions, algorithms and automation predict the end of human
bias and unequal treatment based on belonging to a particular category (absence
of human error), but at the same time, there is a fear of what automation may
mean upfront, i.e. embedded inequality and engendering, which will allow social
antagonisms to be automated. Artificial intelligence or algorithms need learning
data, engendered learning data reproduces and automatises existing process of
engendering. By grouping individuals into categories, quantitative datafication
maintains the relevance of engendered categories of male and female, conse-
quently preserving the reality we see.
Key words: gendering, performativity, artificial intelligence
13
dering. A current example is algorithmic personalization or how artificial intel-
ligence can also automatise the processes of engendering. Ubiquitous mediati-
zation and algorithmically-conditioned visibility reproduce engendering. Mod-
ern datafication classifies individuals into statistical categories, algorithmic rec-
ommendations discourage individuals from exploring alternative solutions. Al-
though personalization is supposed to improve the decision-making process, this
algorithmic relevance relates more to the statistical affiliation of the individual to
the category and is not a reflection of his or her own preferences. According to
some scientific opinions, algorithms and automation predict the end of human
bias and unequal treatment based on belonging to a particular category (absence
of human error), but at the same time, there is a fear of what automation may
mean upfront, i.e. embedded inequality and engendering, which will allow social
antagonisms to be automated. Artificial intelligence or algorithms need learning
data, engendered learning data reproduces and automatises existing process of
engendering. By grouping individuals into categories, quantitative datafication
maintains the relevance of engendered categories of male and female, conse-
quently preserving the reality we see.
Key words: gendering, performativity, artificial intelligence
13
