AIKG-SD 2025 Summe School co-loca ed wi h he NFDI4DS Con e ence 2025
No embe 25-26, 2025, Be lin, Ge many
1
Bias as a signal: a cybe ne ic iew on a ge s a es in AI
Anna Eh enbe g
Abs ac
In a cybe ne ic amewo k o AI-socie y in e ac ions, bias can be unde s ood as a de ia ion
be ween a sys em’s cu en s a e and he a ge s a e i is designed o each. This pos e
explo es he ques ion: How can such a ge s a es be p ac ically de ined o ool-speci ic and
demog aphically di e se use g oups? Th ough a syn hesis o li e a u e and concep ual
inqui y, we map he epis emological, echnical, and e hical challenges in ol ed in de ining
and nego ia ing a ge s a es wi hin AI sys ems.
Cybe ne ics p o ides a way o model AI de elopmen and socie al impac as in e linked
eedback loops. This pe spec i e makes isible how sys em goals eme ge no as objec i e
ac s bu as socially cons uc ed choices shaped by de elope assump ions, ins i u ional
incen i es, egula o y p essu es, and his o ical social s uc u es. As AI sys ems each b oade
audiences and globaliza ion p og esses, he di icul y o de ining sha ed a ge s a es g ows,
especially in he p esence o con lic ing alues such as e iciency, equi y, and au onomy.
The pos e ou lines se e al un esol ed ensions. Epis emological challenges a ise om
si ua ed knowledge: dominan social g oups in luence wha is pe cei ed as a p oblem and
whose expe iences a e e lec ed in sys em design. Technical challenges include ansla ing
complex alues in o measu able pa ame e s, elying on impe ec p oxies, and balancing
anspa ency wi h compe i i e cons ain s. E hical challenges su ace a ound con ex -se ing,
p io i izing pa icula use g oups, and de e mining whose needs and u u es a e ep esen ed
in design choices. These challenges show how a emp s o de ine “desi able” sys em s a es a e
in e wined wi h powe , cul u e, and s uc u al inequi ies.
Ra he han o e ing de ini i e solu ions, his pos e ames bias as a signal wi hin a
cybe ne ic sys em. I shows an indica ion o misalignmen be ween chosen goals and he li ed
eali ies o a ec ed communi ies. By o eg ounding hese un esol ed ensions, he pos e aims
o suppo mo e nuanced discussions abou how a ge s a es in AI a e cons uc ed, con es ed,
and con inuously eshaped h ough eedback be ween echnology and socie y.
Re e ences
Ande son, T. and Shephe d, P. (2025) ‘Modeling and simula ion o human socie ies: om
challenges o new oppo uni ies’, SIMULATION, 101(6), pp. 623–623. A ailable a :
h ps://doi.o g/10.1177/00375497251343097.
A o a, P. and Chowdhu y, R. (2021) ‘C oss-Cul u al Feminis Technologies’, Global
Pe spec i es, 2(1), p. 25207. A ailable a : h ps://doi.o g/10.1525/gp.2021.25207.
Ashby, W.R. (1956) ‘An in oduc ion o cybe ne ics’. A ailable a :
h ps://philpape s.o g/a chi e/ASHAIT (Accessed: 23 Oc obe 2025).
Azzopa di, L. (2021) ‘Cogni i e Biases in Sea ch: A Re iew and Re lec ion o Cogni i e
Biases in In o ma ion Re ie al’, in P oceedings o he 2021 Con e ence on Human
AIKG-SD 2025 Summe School co-loca ed wi h he NFDI4DS Con e ence 2025
No embe 25-26, 2025, Be lin, Ge many
2
In o ma ion In e ac ion and Re ie al. CHIIR ’21: ACM SIGIR Con e ence on Human
In o ma ion In e ac ion and Re ie al, Canbe a ACT Aus alia: ACM, pp. 27–37.
A ailable a : h ps://doi.o g/10.1145/3406522.3446023.
Bee , S. (2002) ‘Wha is cybe ne ics?’, Kybe ne es, 31(2), pp. 209–219. A ailable a :
h ps://doi.o g/10.1108/03684920210417283.
Cody, T., Adams, S. and Beling, P. (2020) ‘Mo i a ing a Sys ems Theo y o AI’, INSIGHT,
23(1), pp. 37–40. A ailable a : h ps://doi.o g/10.1002/ins .12283.
D’Ignazio, C. and Klein, L.F. (2020) Da a Feminism. The MIT P ess. A ailable a :
h ps://doi.o g/10.7551/mi p ess/11805.001.0001.
D age, E., McIne ney, K. and B owne, J. (2024) ‘Enginee s on esponsibili y: eminis
app oaches o who’s esponsible o e hical AI’, E hics and In o ma ion Technology,
26(1), p. 4. A ailable a : h ps://doi.o g/10.1007/s10676-023-09739-1.
Feminis In e ne and Young, J. (2020) ‘Feminis Design Tool De ensible Decision Making
o In e ac ion Design and AI’. A ailable a :
h ps://ugc. u u elea n.com/uploads/ iles/16/b0/16b088ad-6145-45eb-b5d8-
3753a41b4b88/2-10_Feminis DesignTool_2.0.pd (Accessed: 12 No embe 2024).
Fe guson, K.E. (2017) ‘Feminis Theo y Today’, Annual Re iew o Poli ical Science, 20(1),
pp. 269–286. A ailable a : h ps://doi.o g/10.1146/annu e -polisci-052715-111648.
Ga g, S. e al. (2019) ‘Coun e ac ual Fai ness in Tex Classi ica ion h ough Robus ness’.
a Xi . A ailable a : h ps://doi.o g/10.48550/a Xi .1809.10610.
Gene o, D. (2024) ‘Feminisms in A i icial In elligence: Au oma ion Tools owa ds a
Feminis Judicia y Re o m in A gen ina and Mexico | Aymu AI’, Feminis AI
[P ep in ]. A ailable a : h ps:// eminis ai.pubpub.o g/pub/z83eyq54/ elease/1
(Accessed: 4 Oc obe 2024).
Gengle , E.J. e al. (2023) ‘Powe Imbalances in Socie y and AI: On he Need o Expand he
Feminis App oach’.
Ha away, D. (1988) ‘Si ua ed Knowledges: The Science Ques ion in Feminism and he
P i ilege o Pa ial Pe spec i e’, Feminis S udies, 14(3), pp. 575–599. A ailable a :
h ps://doi.o g/10.2307/3178066.
Hayles, N.K. (2023) ‘Technosymbiosis: Figu ing (Ou ) Ou Rela ions o AI’, in J. B owne e
al. (eds) Feminis AI: C i ical Pe spec i es on Algo i hms, Da a, and In elligen
Machines. Ox o d Uni e si y P ess, p. 0. A ailable a :
h ps://doi.o g/10.1093/oso/9780192889898.003.0001.
Heylighen, F. and Joslyn, C. (2003) ‘Cybe ne ics and Second-O de Cybe ne ics’, in
Encyclopedia o Physical Science and Technology. Else ie , pp. 155–169. A ailable
a : h ps://doi.o g/10.1016/B0-12-227410-5/00161-7.
Klein, L. and D’Ignazio, C. (2024) ‘Da a Feminism o AI’, in The 2024 ACM Con e ence on
Fai ness, Accoun abili y, and T anspa ency. FAccT ’24: The 2024 ACM Con e ence
on Fai ness, Accoun abili y, and T anspa ency, Rio de Janei o B azil: ACM, pp. 100–
112. A ailable a : h ps://doi.o g/10.1145/3630106.3658543.
Mühlho , R. (2025) ‘AI Cybe ne ics’, in The E hics o AI. B is ol Uni e si y P ess, pp. 135–
145. A ailable a : h ps://b is oluni e si yp essdigi al.com/monochap-
oa/book/9781529249262/ch008.xml (Accessed: 23 Oc obe 2025).
AIKG-SD 2025 Summe School co-loca ed wi h he NFDI4DS Con e ence 2025
No embe 25-26, 2025, Be lin, Ge many
3
Polcumpally, A.T. (2022) ‘A i icial in elligence and global powe s uc u e: unde s anding
h ough Luhmann’s sys ems heo y’, AI & SOCIETY, 37(4), pp. 1487–1503. A ailable
a : h ps://doi.o g/10.1007/s00146-021-01219-8.
Pu kayas ha, B. (2012) ‘In e sec ionali y in a T ansna ional Wo ld’, Gende & Socie y, 26(1),
pp. 55–66. A ailable a : h ps://doi.o g/10.1177/0891243211426725.
Radanlie , P. (2025) ‘AI E hics: In eg a ing T anspa ency, Fai ness, and P i acy in AI
De elopmen ’, Applied A i icial In elligence, 39(1), p. 2463722. A ailable a :
h ps://doi.o g/10.1080/08839514.2025.2463722.
S i as a a, A. e al. (2023) ‘Beyond he Imi a ion Game: Quan i ying and ex apola ing he
capabili ies o language models’. a Xi . A ailable a :
h ps://doi.o g/10.48550/a Xi .2206.04615.
S een, M. e al. (2025) ‘We need be e images o AI and be e con e sa ions abou AI’, AI &
SOCIETY, 40(5), pp. 3615–3626. A ailable a : h ps://doi.o g/10.1007/s00146-024-
02101-z.
Toupin, S. (2024) ‘Shaping eminis a i icial in elligence’, New Media & Socie y, 26(1), pp.
580–595. A ailable a : h ps://doi.o g/10.1177/14614448221150776.
Wudel, Alexand a (2023) Na iga ing he In e sec ion: Disin o ma ion, Deep akes and
Gende -Based Violence, FemAI. A ailable a : h ps://www. em-
ai.com/pos /na iga ing- he-in e sec ion-deep akes-disin o ma ion-and-gende -based-
iolence (Accessed: 10 Oc obe 2024).
