Decoding Bias in Gene a i e AI. F aming Socio-Technical
Da a Li e acy as a Collec i e C i ical P ac ice
An onella Au uo i1,2
1RMIT Uni e si y, School o Design, Melbou ne, Aus alia
2Ins i u e o Design, SUPSI Uni e si y o Applied Sciences and A s o Sou he n Swi ze land, Mend isio, Swi ze land
Abs ac
This p ac ice-based PhD de elops a socio- echnical da a li e acy amewo k o gene a i e AI, emphasizing
pa icipa o y engagemen wi h he socio-cul u al dimensions o da a. Th ough me hods such as pa icipa o y
wo kshops and c i ical making, he esea ch demons a es how non- echnical s akeholde s can decode and
in e ene in algo i hmic bias h ough engagemen wi h classi ica ion da a p ac ices. The esul ing oolki and
e alua i e amewo k o e p ac ical s a egies o inclusi e, cul u ally-awa e pa icipa ion in educa ional and
ci ic con ex s. By concep ualizing da a li e acy as a c i ical, si ua ed and collec i e c i ical p ac ice, he esea ch
con ibu es o HCI by ad ancing mo e equi able and ela ional human-AI in e ac ion.
Keywo ds
gene a i e a i icial in elligence, bias, human-AI in e ac ion, da a li e acy, eminis epis emologies
1. In oduc ion
Gene a i e AI echnologies a e inc easingly cen al o he in as uc u es ha media e pe cep ion,
ep esen a ion, and decision-making; howe e , hese sys ems also eplica e and in ensi y exis ing social
hie a chies h ough his o ically embedded biases in aining da a and classi ica ion a chi ec u es [
1
,
2
,
3
].
This p ac ice-based doc o al esea ch in oduces socio- echnical da a li e acy as a c i ical and e lexi e
amewo k o engaging wi h hese sys ems, one ha equips non- echnical use s o decode, in e oga e,
and in e ene in he classi ica o y logics ha shape gene a i e isual ou pu s. Socio- echnical da a
li e acy is de ined he e as a si ua ed capaci y o c i ically engage wi h he in e connec ed echnical and
socio-cul u al dimensions o gene a i e AI.
The e m socio- echnical unde sco es ha hese sys ems a e no pu ely compu a ional bu a e shaped
by cul u al assump ions, ins i u ional s uc u es, and powe ela ions. Da a e e s no only o aining
co po a bu also o he classi ica ion schemas ha o ganize and gi e meaning o in o ma ion. These
classi ica ions—o en hidden behind polished ou pu s—play a c ucial ole in de e mining wha is made
isible, no ma i e, o excluded. Li e acy, in his con ex , is no me ely a echnical skill bu a ela ional,
c i ical abili y o in e p e , ques ion, and eshape algo i hmic ep esen a ions in a si ua ed con ex . I
enables use s o su ace bias, nego ia e meaning, and collec i ely eimagine he epis emic s uc u es
embedded in gene a i e sys ems.
Ra he han ea ing use s as passi e ecipien s o gene a i e echnologies, his wo k emphasizes
hei ole as epis emic agen s wi h e hical esponsibili y in shaping model beha io h ough in e ac ions
such as p omp design, con en selec ion, and sense-making. Wi hin his pe spec i e, he esea ch
in es iga es how non- echnical s akeholde s can ac i ely challenge dominan ep esen a ions and
co-cons uc al e na i e classi ica o y logics wi hin gene a i e AI sys ems.
The wo k is guided by wo cen al ques ions: How can socio- echnical da a li e acy unc ion as
a c ea i e-c i ical me hod o decoding and in e ening in bias wi h gene a i e AI sys ems, while
suppo ing use agency in he in e p e a ion and manipula ion o classi ica ion p ocesses? How can
CHI aly 2025: Technologies and Me hodologies o Human-Compu e In e ac ion in he Thi d Millenium, Doc o al Conso ium,
6-10 Oc obe 2025, Sale no, I aly
$[email p o ec ed] (A. Au uo i)
0000-0002-5725-8446 (A. Au uo i)
©2025 Copy igh o his pape by i s au ho s. Use pe mi ed unde C ea i e Commons License A ibu ion 4.0 In e na ional (CC BY 4.0).
pa icipa o y app oaches o da a and AI mo e beyond echnical pe o ma i i y o os e mo e ela ional,
ca e-o ien ed engagemen s wi h gene a i e echnologies and da a?
In o med by eminis epis emologies [
4
,
5
], c i ical pedagogy [
6
,
7
], and c i ical da a s udies [
8
,
9
],
his esea ch app oaches knowledge as si ua ed, ela ional, and ma e ially embedded. Feminis heo y
ejec s he no ion o disembodied objec i i y, emphasizing ins ead pa ial pe spec i es g ounded in li ed
expe ience and condi ioned by speci ic socio-cul u al con ex s. This lens suppo s an unde s anding
o human–AI in e ac ion as media ed by bo h pe sonal and s uc u al ac o s, such as iden i y, a ec ,
memo y, language, and access. C i ical pedagogy ein o ces his pe spec i e by aming lea ning as
a dialogic and collec i e p ocess, o ien ed owa d e lec ion, agency, and he dis up ion o dominan
epis emic hie a chies. C i ical da a s udies u he ex end his app oach by in e oga ing how da a
in as uc u es embed social assump ions and ep oduce asymme ies o ep esen a ion. In addi ion
o his heo e ical ounda ion, specula i e design [
10
] can o e me hodological s a egies ha enable
pa icipan s o challenge no ma i e da a logics and imagine al e na i e engagemen s wi h gene a i e
echnologies.
Si ua ed wi hin HCI, his esea ch ad ances a se o ools designed o e alua e and in o m non-
echnical use s’ engagemen wi h gene a i e AI sys ems, wi h a ocus on os e ing agency and c i ical
awa eness. As hese sys ems become inc easingly embedded in e e yday con ex s, unde s anding how
use s in e p e , challenge, and in luence gene a i e ou pu s becomes essen ial.
Designed o educa ional and ci ic se ings, hese ools suppo e lec i e engagemen and p omo e
indi idual and collec i e esponsibili y. By o eg ounding use agency in shaping model beha io , he
amewo k in i es c i ical a en ion o accoun abili y, anspa ency, and he classi ica o y sys ems ha
unde pin algo i hmic ou pu s. I hus expands he scope o da a li e acy owa d mo e equi able, si ua ed,
and e hically esponsi e human–AI ela ions.
This esea ch is cu en ly a he beginning o i s second yea , wi h he me hodology de ined and
unde implemen a ion. E hics app o al has been ob ained om he RMIT Uni e si y Human Resea ch
E hics Commi ee in Melbou ne, and expe in e iews and pa icipa o y wo kshops a e cu en ly
unde way. The ollowing sec ions ou line he heo e ical ounda ions, pa icipa o y me hodology, and
p ac ical con ibu ions o his esea ch.
2. Resea ch Backg ound
The on ological ounda ion o his p ac ice-based PhD engages wi h in e disciplina y discou ses om
philosophy, his o y, and science and echnology s udies o examine how echnological sys ems, bias,
and disc imina ion a e co-p oduced. Unde s anding AI as a ma e ial and discu si e in as uc u e—
shaped by p ac ices o classi ica ion, ep esen a ion, and decision-making— equi es a his o ically and
concep ually g ounded pe spec i e [2,11].
Recen esea ches demons a e ha bo h la ge language models (LLMs) and ex - o-image (TTI)
sys ems sys ema ically ep oduce and ampli y s e eo ypes ela ed o gende , sexuali y, and e hnici y
[
12
,
13
,
14
], ma ginalize non-Wes e n epis emologies [
15
], and cons ain iden i y ep esen a ions [
16
],
while ul ima ely excluding disabili y and neu odi e gen iden i ies [
17
]. These pa e ns o exclusion a e
deeply embedded in he cons uc ion o aining da ase s, whe e selec ion, anno a ion, and classi ica ion
p ac ices a e shaped by assump ions abou wha and who should be made isible. Such p ocesses con e
epis emic legi imacy on pa icula wo ld iews, embedding hem wi hin algo i hmic sys ems unde he
guise o echnical objec i i y.
Da a anno a ion has been shown o be a si ua ed and powe -laden p ocess ha media es subjec i i y
and ins i u ional au ho i y [
18
,
19
]. Visual axonomies, such as hose ound in ImageNe , ha e been
shown o ely on no ma i e assump ions abou wha iden i ies, oles, and exp essions should look like,
educing indi iduals o p ede ined labels ha claim uni e sali y bu e lec na ow, cul u ally speci ic
wo ld iews [
3
]. This educ i e logic p esumes a di ec , s able co espondence be ween appea ance and
meaning, ein o cing cul u al s e eo ypes unde he guise o compu a ional legibili y. ImageNe , in
pa icula , exempli ies he isks o la ge-scale anno a ion when applied o human subjec s, whe e labels
such as “lose ,” “klep omaniac,” o “sla e n” we e a ached o images o eal people wi hou consen o
con ex ual nuance [3].
As a gued by Bowke and S a , classi ica ion sys ems embody ins i u ional and cul u al logics ha ,
al hough o en obscu ed, ca y signi ican epis emic and poli ical consequences [
20
]. In his sense,
da ase s do no me ely e lec eali y, bu ac i ely cons uc i h ough he wo ld iews embedded in
hei s uc u es and labeling p ac ices [
21
]. The ac o selec ing, labeling, and ca ego izing images is no
neu al o echnical—i is a poli ical in e en ion wi h las ing impac on how people a e seen, so ed,
and ac ed upon by AI sys ems. These classi ica o y egimes no only ep oduce ha m bu ha e become
inc easingly opaque as comme cial AI sys ems scale, limi ing public sc u iny o how ep esen a ions
a e p oduced and deployed.
This esea ch con ibu es o ongoing deba es by aming da a li e acy as a c i ical, collec i e p ac ice
speci ic o he con ex o gene a i e AI. I add esses he epis emic and poli ical dimensions o clas-
si ica ion sys ems, p oposing ools and me hods ha make hese s uc u es accessible and open o
con es a ion by non- echnical audiences.
3. Towa ds a Bo om-Up Pa icipa o y Impe a i e
A cen al ambi ion o his esea ch is o in ol e non- echnical expe s, communi y membe s, educa o s,
ac i is s, and o he s akeholde s who ha e been his o ically ma ginalized o excluded om he design
and classi ica ion p ocesses unde pinning AI sys ems.
These indi iduals a e no me ely end-use s; hey a e co-cons uc o s o knowledge [
22
,
23
] whose
li ed expe iences, alues, and pe spec i es a e essen ial o su acing biases, con es ing dominan
na a i es, and en isioning al e na i e u u es o AI [24].
Con en ional pa icipa o y p ac ices in AI o en con ine s akeholde in ol emen o consul a i e o
okenis ic oles, whe e inpu is solici ed only a disc e e s ages— ypically a e key design decisions ha e
al eady been made—o es ic ed o supe icial aspec s such as use in e ace eedback [
25
]. Mo eo e ,
cu en p ac ice equen ly elies on p oxies—such as UX p o essionals o algo i hmic models— o
ep esen s akeholde oices, a he han enabling di ec and sus ained engagemen [26].
This esul s in cons ained agency and limi ed in luence o e ounda ional classi ica o y s uc u es.
Such o ms o engagemen a e no me ely desi able bu cons i u e necessa y condi ions o ensu ing
accoun abili y, anspa ency, and con ex ual ele ance in AI sys ems. Wi hin his amewo k, he
edis ibu ion o epis emic au ho i y and he collec i e shaping o classi ica ion p ocesses a e unde s ood
as cen al o ad ancing mo e jus , e lexi e, and socially esponsi e echnological u u es [
27
,
28
,
26
,
29
].
This equi es poli ically in o med unde s andings o how echnology and ci izenship a e en angled,
making isible he powe ela ions embedded in digi al sys ems and suppo ing emancipa o y p ac ices
aimed a social jus ice[
30
]. As D’Ignazio and Klein emphasize in hei amewo k o da a eminism,
Expanding who pa icipa es in he design and in e p e a ion o da a is no simply a ma e o b oadening
access, bu a delibe a e epis emic choice—one ha challenges dominan knowledge sys ems and a i ms
he alue o si ua ed, plu al o ms o unde s anding ha a e o en excluded om mains eam da a
p ac ices [31].
4. Me hodology
This PhD esea ch adop s a p ac ice-based me hodology ha wea es oge he c i ical heo e ical
inqui y and pa icipa o y expe imen a ion. The app oach is s uc u ed a ound ou in e connec ed
me hodological pilla s, each designed o o eg ound he epis emic and e hical complexi ies o gene a i e
AI while cen e ing he agency o non- echnical s akeholde s.
4.1. Li e a u e Re iew as C i ical In as uc u e Mapping
The i s pilla consis s o a li e a u e e iew concei ed as a o m o c i ical in as uc u e mapping.
Ra he han summa izing exis ing wo k, i delinea es he concep ual landscape o da a classi ica ion,
ep esen a ional bias, and use agency wi hin gene a i e AI. These sys ems a e app oached as socio-
echnical in as uc u es shaped by his o ical, cul u al, and poli ical condi ions.
As pa o his phase, a c oss-disciplina y bias ca og aphy is assembled—d awing om media
s udies, da a science, in o ma ics, and psychology—no o isola e echnological ailu es, bu o examine
how algo i hmic and human biases in e sec . This mapping suppo s a ela ional unde s anding o
classi ica ion p ocesses and p o ides a sha ed ounda ion o e lec ion and discussion wi h expe s and
pa icipan s in ollowing me hods.
4.2. Expe In e iews
The second pilla in ol es semi-s uc u ed in e iews wi h designe s, educa o s, AI p ac i ione s, and
ac i is s engaged in c i ical wo k on algo i hmic sys ems. Concei ed as dialogical a he han ex ac i e,
hese in e iews suppo knowledge co-p oduc ion h ough e lec i e p omp s and in-dep h discussion.
Pa icipan s a e in i ed o c i ically engage wi h and con ibu e o he c oss-disciplina y bias ca -
og aphy de eloped du ing he li e a u e e iew, b inging insigh s om hei espec i e domains o
p ac ice. This p ocess su aces ensions be ween e hical commi men s and echnical cons ain s, and
in o ms he i e a i e co-design o wo kshop o ma s and he de elopmen o e alua ion c i e ia.
4.3. Pa icipa o y Wo kshops
The hi d me hodological pilla cen e s on pa icipa o y wo kshops, which a e concei ed as epis emic
in e en ions and c i ical making spaces. These wo kshops in i e pa icipan s o engage wi h gene a i e
AI h ough a combina ion o e lec i e inqui y and specula i e expe imen a ion.
Ac i i ies include p omp hacking—whe e pa icipan s i e a i ely es and anno a e gene a i e models
o e eal hidden biases and zine-making, which d aws on eminis adi ions o s o y elling o documen
pe sonal encoun e s wi h algo i hmic classi ica ion. Fu he , da ase emixing enables he co-c ea ion o
al e na i e axonomies h ough collabo a i e da a cu a ion, while image anno a ion and eclassi ica ion
exe cises encou age pa icipan s o challenge no ma i e isual g amma s and dis up es ablished
hie a chies.
These wo kshops a e guided by p inciples o cons uc ionism [
32
], ca e e hics [
33
], and pedagogical
co-p oduc ion [
6
], emphasizing hands-on, embodied c i ique o e abs ac delibe a ion. Da a gene a ed
om obse a ions, pa icipan -c ea ed a i ac s, and e lec i e discussions a e analyzed using e lexi e
hema ic analysis.
4.4. Toolki and E alua i e F amewo k o Socio-Technical Da a Li e acy in
Pa icipa o y and Educa ional En i onmen
The inal phase o he esea ch in ol es he de elopmen and assessmen o a modula oolki designed
o suppo he p ac ice o socio- echnical da a li e acy in educa ional and ci ic con ex s. The oolki
includes adap able acili a ion o ma s, design p obes, and guiding ma e ials, and is examined h ough
an e alua i e lens ha a ends o mul iple dimensions o pa icipan engagemen .
These include epis emic unde s anding— he abili y o a icula e and iden i y bias in gene a i e AI
sys ems; social engagemen — e lec ed in he willingness o discuss and in e ene in classi ica ion
p ocesses; and c i ical awa eness— he ecogni ion o AI sys ems as si ua ed, alue-laden in as uc u es.
E alua ion is no ea ed as a conclusi e s ep, bu a he as an i e a i e and gene a i e momen wi hin
he esea ch, one ha e lec s on how heo e ical commi men s a e ansla ed in o si ua ed p ac ice, and
how collec i e inqui y can in o m mo e inclusi e and con ex ually g ounded ways o engaging wi h AI.
5. Conclusion and Resea ch Con ibu ion
This p ac ice-based doc o al esea ch con ibu es o HCI by e aming AI li e acy h ough a socio-
echnical lens ha cen e s in e p e i e agency, e hical esponsibili y, and pa icipa o y engagemen .
Gene a i e AI is app oached no as a neu al in as uc u e bu as a si e whe e bias is embedded, enac ed,
and con es ed h ough in e ac ion.
The de elopmen o a socio- echnical da a li e acy amewo k enables non- echnical s akeholde s—
o en excluded om p ocesses o design, go e nance, and in e p e a ion— o c i ically engage wi h he
classi ica o y logics shaping gene a i e ou pu s. The p ojec in es iga es how engagemen wi h gene a-
i e AI can suppo a shi om bias awa eness o he cul i a ion o e hical agency. I ocuses on how
indi iduals ecognize and espond o algo i hmic ep esen a ions, and how momen s o in e p e a ion
can become si es o nego ia ion, e lexi i y, and sha ed accoun abili y. Bias is unde s ood no only as
a p ope y o da a bu as some hing ep oduced h ough in e ac ion, sense-making, and ins i u ional
con ex .
Th ough pa icipa o y wo kshops, dialogical in e iews, and c i ical making p ac ices, he esea ch
explo es how collec i e, si ua ed in e en ions can os e mo e inclusi e and plu alis ic app oaches
o knowledge p oduc ion. The esul ing oolki and e alua i e amewo k o e p ac ical esou ces
o educa ional and ci ic con ex s, while p oposing new ways o assessing agency, awa eness, and
engagemen in AI-media ed en i onmen s. Ul ima ely, his wo k expands he scope o da a li e acy and
pa icipa o y design owa d mo e equi able, e lexi e, and socially esponsi e human–AI in e ac ions.
Acknowledgmen s
The au ho would like o hank hei supe iso s, B ad Haylock and Lau ene Vaughan, om he RMIT
Design School o hei unwa e ing suppo and o guiding he de elopmen o his esea ch.
Decla a ion on Gene a i e AI
Du ing he p epa a ion o his wo k, he au ho used G amma ly in o de o: G amma and spelling
check. Fu he , he au ho used Deeply in o de o: Syn ax e iew.
Re e ences
[1]
M. B oussa d, Mo e han a Gli ch: Con on ing Race, Gende , and Abili y Bias in Tech, MIT P ess,
2023.
[2]
K. C aw o d, A las o AI: Powe , poli ics, and he plane a y cos s o a i icial in elligence, Yale
Uni e si y P ess, 2021.
[3]
K. C aw o d, T. Paglen, Exca a ing ai: The poli ics o images in machine lea ning aining se s,
h ps://exca a ing.ai/, 2020.
[4]
D. Ha away, Si ua ed knowledges: The science ques ion in eminism and he p i ilege o pa ial
pe spec i e, Feminis S udies 14 (1988). doi:10.2307/3178066.
[5]
S. Ha ding, Subjec i i y, expe ience and knowledge: An epis emology om/ o ainbow coali ion
poli ics, De elopmen and Change 23 (1992).
[6] P. F ei e, Pedagogy o he opp essed, Seabu y P ess, 1970.
[7] B. Hooks, Teaching o ansg ess: Educa ion as he p ac ice o eedom, Rou ledge, 1994.
[8]
J. G ay, C. Ge li z, L. Bouneg u, Da a in as uc u e li e acy, Big Da a & Socie y 5 (2018).
doi:10.1177/2053951718786316.
[9]
L. Pang azio, N. Selwyn, C i ical Da a Li e acies: Re hinking Da a and E e yday Li e, MIT P ess,
2023.
[10]
N. Sánchez Que ubín, S. Niede e , Clima e u u es: machine lea ning om cli- i, Con e gence 30
(2024). doi:10.1177/13548565221135715.
[11] G. Simondon, Du mode d’exis ence des obje s echniques, Aubie , 1958.
[12]
A. S. Luccioni, C. Akiki, M. Mi chell, Y. Je ni e, S able bias: Analyzing socie al ep esen a ions in
di usion models, a Xi :2303.11408, h ps://a xi .o g/abs/2303.11408, 2023.
[13]
L. Nicole i, D. Bass, Humans a e biased. gene a i e ai is e en wo se, Bloombe g Technology +
Equali y, h ps://www.bloombe g.com/g aphics/2023-gene a i e-ai-bias, 2023.
[14]
UNESCO, Gene a i e ai: Unesco s udy e eals ala ming e idence
o eg essi e gende s e eo ypes, h ps://www.unesco.o g/en/a icles/
gene a i e-ai-unesco-s udy- e eals-ala ming-e idence- eg essi e-gende -s e eo ypes, 2024.
[15]
V. Tu k, How ai educes he wo ld o s e eo ypes, Res o Wo ld, h ps:// es o wo ld.o g/2023/
ai-image-s e eo ypes/, 2023.
[16]
A. F. d. C. Vázquez, E. C. Ga ido-Me chán, A axonomy o he biases o he images c ea ed by
gene a i e a i icial in elligence, a Xi :2407.01556, h ps://doi.o g/10.48550/a Xi .2407.01556, 2024.
[17]
Y. Welke , Algo i hmic di e si y: Mi iga ing ai bias and disabili y exclu-
sion, Fo bes, h ps://www. o bes.com/councils/ o bes echcouncil/2023/05/09/
algo i hmic-di e si y-mi iga ing-ai-bias-and-disabili y-exclusion/, 2023.
[18]
M. Miceli, M. Schuessle , T. Yang, Be ween subjec i i y and imposi ion: Powe dynamics in
da a anno a ion o compu e ision, P oc. ACM Hum.-Compu . In e ac . 4 (2020). doi:
10.1145/
3415186.
[19]
A. A zbe ge , S. Buijsman, M. L. Lupe i, A. Bozzon, J. Yang, No hing comes wi hou i s wo ld –
p ac ical challenges o aligning llms o si ua ed human alues h ough lh , in: P oceedings o he
2024 AAAI/ACM Con e ence on AI, E hics, and Socie y, AIES ’24, AAAI P ess, San Jose, CA, USA,
2024. doi:10.1609/aies. 7i1.31617.
[20] G. C. Bowke , S. L. S a , So ing hings ou : Classi ica ion and i s consequences, MIT P ess, 1999.
[21]
H. Da is, A da ase is a wo ld iew, Medium, h ps://medium.com/da a-science/
a-da ase -is-a-wo ld iew-5328216dd44d, 2019.
[22]
C. DiSal o, Design and he cons uc ion o publics, Design Issues 25 (2009). doi:
10.1162/desi.
2009.25.1.48.
[23]
A. K. Munk, V. Tommaso, A. Meunie , Da a sp in s: A collabo a i e o ma in digi al con o e sy
mapping, in: J. Ve esi, D. Ribes (Eds.), Digi al STS: A ield guide o science & echnology s udies,
P ince on Uni e si y P ess, 2019. doi:10.2307/j.c c77mp9.34.
[24]
S. Cos anza-Chock, Design jus ice: Communi y-led p ac ices o build he wo lds we need, MIT
P ess, 2020.
[25]
F. Delgado, S. Yang, M. Madaio, Q. Yang, The pa icipa o y u n in ai design: Theo e ical ounda ions
and he cu en s a e o p ac ice, in: EAAMO ’23: P oceedings o he 3 d ACM Con e ence
on Equi y and Access in Algo i hms, Mechanisms, and Op imiza ion, ACM, 2023. doi:
h ps:
//doi.o g/10.1145/3617694.3623261.
[26]
M. Sloane, E. Moss, O. Awomolo, L. Fo lano, Pa icipa ion is no a design ix o machine lea ning,
in: P oceedings o he 2nd ACM Con e ence on Equi y and Access in Algo i hms, Mechanisms, and
Op imiza ion EAAMO ’22:, ACM, 2020. doi:h ps://doi.o g/10.1145/3551624.3555285.
[27]
HEK House o Elec onic A s Basel, Pa icipa o y ai: How o make be e ai, h ps://sha e.hek.ch/
en/pa icipa o y-ai-how- o-make-be e -ai, 2024.
[28]
D. J. Nuce a, V. Mogile ich, Teaching Communi y Technology Handbook, De oi Communi y
Technology P ojec , 2015.
[29]
P. Gou le , D. Ricci, M. C épel, Reclaiming a i icial in elligence accoun s: A plea o a pa icipa o y
u n in a i icial in elligence inqui ies, Big Da a & Socie y Ap il-June (2024). doi:
10.1177/
20539517241248093.
[30]
A. Emejulu, C. McG ego , Towa ds a adical digi al ci izenship in digi al educa ion, C i ical S udies
in Educa ion 60 (2019). doi:10.1080/17508487.2016.1234494.
[31] C. D’Ignazio, L. F. Klein, Da a eminism, MIT P ess, 2020.
[32] S. Pape , Minds o ms—Child en, Compu e s and Powe ul Ideas, Basic Books, 1980.
[33]
M. Puig de la Bellacasa, Ma e s o ca e in echnoscience: Assembling neglec ed hings, Social
S udies o Science 41 (2017). doi:10.1177/0306312710380301.
