Taylo , Maisy; Vollme , Sa ah; Ra a , Zaynab; Benjamin, Ga ield
A icle
Who's a s ake? The (non)pe o ma i i y o "s akeholde s"
in UK ech policy
In e ne Policy Re iew
P o ided in Coope a ion wi h:
Alexande on Humbold Ins i u e o In e ne and Socie y (HIIG), Be lin
Sugges ed Ci a ion: Taylo , Maisy; Vollme , Sa ah; Ra a , Zaynab; Benjamin, Ga ield (2025) : Who's
a s ake? The (non)pe o ma i i y o "s akeholde s" in UK ech policy, In e ne Policy Re iew, ISSN
2197-6775, Alexande on Humbold Ins i u e o In e ne and Socie y, Be lin, Vol. 14, Iss. 3, pp.
1-31,
h ps://doi.o g/10.14763/2025.3.2033
This Ve sion is a ailable a :
h ps://hdl.handle.ne /10419/330352
S anda d-Nu zungsbedingungen:
Die Dokumen e au EconS o dü en zu eigenen wissenscha lichen
Zwecken und zum P i a geb auch gespeiche und kopie we den.
Sie dü en die Dokumen e nich ü ö en liche ode komme zielle
Zwecke e iel äl igen, ö en lich auss ellen, ö en lich zugänglich
machen, e eiben ode ande wei ig nu zen.
So e n die Ve asse die Dokumen e un e Open-Con en -Lizenzen
(insbesonde e CC-Lizenzen) zu Ve ügung ges ell haben soll en,
gel en abweichend on diesen Nu zungsbedingungen die in de do
genann en Lizenz gewäh en Nu zungs ech e.
Te ms o use:
Documen s in EconS o may be sa ed and copied o you pe sonal
and schola ly pu poses.
You a e no o copy documen s o public o comme cial pu poses, o
exhibi he documen s publicly, o make hem publicly a ailable on he
in e ne , o o dis ibu e o o he wise use he documen s in public.
I he documen s ha e been made a ailable unde an Open Con en
Licence (especially C ea i e Commons Licences), you may exe cise
u he usage igh s as speci ied in he indica ed licence.
h ps://c ea i ecommons.o g/licenses/by/3.0/de/deed.en
Volume 14 |
Who’s a s ake? The (non)pe o ma i i y o
“s akeholde s” in UK ech policy
Maisy Taylo Uni e si y o Yo k
Sa ah Vollme Uni e si y o U ech
Zaynab Ra a Solen Uni e si y
Ga ield Benjamin Uni e si y o Camb idge
DOI: h ps://doi.o g/10.14763/2025.3.2033
Published: 4 Sep embe 2025
Recei ed: 30 Janua y 2024 Accep ed: 15 No embe 2024
Funding: The au ho s would like o acknowledge unding ecei ed om Solen
Uni e si y o his esea ch.
Compe ing In e es s: The au ho has decla ed ha no compe ing in e es s exis ha
ha e in luenced he ex .
Licence: This is an open-access a icle dis ibu ed unde he e ms o he C ea i e
Commons A ibu ion 3.0 License (Ge many) which pe mi s un es ic ed use,
dis ibu ion, and ep oduc ion in any medium, p o ided he o iginal wo k is p ope ly
ci ed. h ps://c ea i ecommons.o g/licenses/by/3.0/de/deed.en
Copy igh emains wi h he au ho (s).
Ci a ion: Taylo , M., Vollme , S., Ra a , Z., & Benjamin, G. (2025). Who’s a s ake? The
(non)pe o ma i i y o “s akeholde s” in UK ech policy. In e ne Policy Re iew, 14(3).
h ps://doi.o g/10.14763/2025.3.2033
Keywo ds: S akeholde , Powe , Pe o ma i i y, Uni ed Kingdom, Tech policy
Abs ac : The e m "s akeholde " ea u es p ominen ly in discou ses su ounding ech policy. I is
used as a ma ke o engaging wi h wide o ganisa ions and publics beyond go e nmen and he
companies ha make echnologies. Bu wha does his e m do in p ac ice? Wha oles does i
c ea e o deny, wha powe s uc u es does i open o close? This s udy o 194 ech policy
documen s p oduced by he UK go e nmen o e a span o i e yea s unco e s he di e en ways
he e m "s akeholde " is used in p ac ice. By compa ing he use o he wo d wi h whose ideas a e
ac ually ci ed in he documen s, i highligh s he discu si e gaps be ween wha is claimed and
wha is ep esen ed in he oices ha shape policy. The esul s a e analysed h ough quee
pe o ma i i y, including ins i u ional non-pe o ma i i y and pe i-pe o ma i e amings, o assess
he oles ha a e imposed on di e en g oups, and he di e en hie a chies and powe s uc u es
he s akeholde cons uc s in cu en UK policy p ac ices.
Issue 3
INTRODUCTION
Tech policy in he UK has been unde going a apid inc ease in ac i i y and p omi-
nence in ecen yea s. Following he adop ion o he Eu opean Gene al Da a P o-
ec ion Regula ion (GDPR) in 2016, and in he con ex o UK law di e ging a e i s
depa u e om he EU, he e has been a lu y o p oposed egula ions, including
mul iple digi al s a egies ( o example, DSIT and DCMS, 2019; DSIT e al., 2021)
and a majo AI whi e pape (DSIT and O ice o AI, 2023). The UK has aced o lead
he wo ld in ackling he escala ing ha ms o echnologies, o example wi h he
Online Sa e y Ac 2023, while p omo ing i s alues and in e es s. This has led o a
weal h o documen s: policy p oposals, consul a ions, guidance, codes and o he s.
These ha e been c ea ed by depa men s and egula o s om ac oss go e nmen ,
om he Depa men o Cul u e, Media and Spo (DCMS) and Depa men o
Business, Ene gy and Indus ial S a egy (BEIS) o he Cen al Digi al and Da a O -
ice and he Na ional Audi O ice, he In o ma ion Commissione ’s O ice (ICO) and
O Com, and he Child en’s Commissione , among many o he s, including an in-
c ease in c oss-depa men al collabo a ions. And ye he UK’s indus ial s a egy-
d i en app oach has lagged behind, o example, he EU and China, in e ms o e -
ec i e egula ion ha mee s he needs o e e yday ci izens, and he e ec s o he
policy enzy emain con es ed, wi h he Online Sa e y Ac , he lagship egula ion
enac ed so a , being labelled by London Mayo Sadiq Khan as “no i o pu pose”
(Elgo and Cou ea, 2024) in esponse o speci ic c ises.
In pa , he di icul ies a e due o he shee b ead h o social and echnical p ob-
lems o conside when a emp ing o manage he wide a ie y o issues wi h
which ech policy mus con end, in e sec ing wi h many di e en o he policy a -
eas om heal h o educa ion o policing o business o ene gy and he en i on-
men . Tech policy is also no able o he con inued con es a ion o key e ms. Much
deba e has been had, o example, a ound he de ini ion o “AI” in legisla ion ( he
EU AI Ac e ised and expanded i s de ini ions and ca ego ies h oughou i s de el-
opmen ), wi h Imogen Pa ke o he Ada Lo elace Ins i u e poin ing ou how lack
o de ini ional cla i y “lea e[s] he public sec o wi hou ac ionable o cohesi e
amewo ks o decision-making” (2025). When he e y hings being egula ed a e
s ill being de ined, i is he e o e a key momen in which o ensu e ha hese
c oss-cu ing and a - eaching policies a e embedding he e ms and alues ha
ep esen he di e en g oups wi h in e es s a s ake in he deba es. This issue o
s akeholde s is impo an . Many documen s claim o conside o ep esen he in-
e es s o “wide- anging s akeholde s”, o o consul wi h s akeholde s. Bu wha
exac ly does “s akeholde ” mean? Despi e who is consul ed, whose in e es s ac u-
2 In e ne Policy Re iew 14(3) | 2025
ally ollow h ough o in luence policy? Who coun s as a s akeholde in di e en
ace s o ech policy? Is s akeholde used o ep esen ideals o he in e es s o
hose whose li es a e a s ake? Who has a s ake in he bene i s, he ha ms, he de-
cisions?
In his esea ch we examined 194 documen s o UK ech policy o unco e how
widely he e m s akeholde is used. We wan ed o look benea h he he o ic o
consul a ion o see wha sou ces ge ci ed in he inal p esen a ion o policies, and
by ex ension whose in e es s a e eally ep esen ed. The p ojec is based on he
ollowing key ques ions:
• How widely is he e m “s akeholde ” employed in UK ech policy?
• Whose iews a e ac ually ep esen ed in policy documen s?
• How does his ep esen a ion di e be ween echnology a eas, issues, o
depa men s?
• Wha does he use o he e m s akeholde c ea e, conceal o exclude?
Answe ing hese ques ions ge s o he hea o wha i means o ha e a s ake in
echnology, and who is a s ake in he way ech policy is de eloped. Pu suing hese
a gumen s highligh s he need o look a exclusion and ma ginalisa ion no only in
a p ac ical sense bu also a he discu si e le el. The use o he e m s akeholde
implies ce ain oles, ye hey emain ill-de ined, malleable and open o being op-
e a ionalised o poli ical ends. S akeholde engagemen could be a ool o
widening pa icipa ion in ech policy, o i could become a eil o assigning powe
o hose wi h exis ing in luence o a inancial s ake in s ee ing egula ion away
om mo e ans o ma i e o edis ibu i e aims. I is impo an o conside ap-
p oaches o analysing hese discou ses and he di e en ways ha he same e m
can be used o c ea e powe o deny i .
The analysis is he e o e amed h ough he lens o pe o ma i i y, b inging quee
eminis p inciples o unde s anding he ways ha powe is embedded in domi-
nan discou ses, oles and beha iou s. Building on he wo k o Judi h Bu le (1990;
2018), we applied he idea ha speech and naming can c ea e ce ain iden i ies
and powe s uc u es when people ake on (o a e o ced o ake on) ce ain oles.
This app oach helps us o unde s and when he labelling o a g oup as ha ing a
s ake in policy ac ually cons i u es ha ole and a o ds in luence o b ings hei
in e es s o he o e. We used Sa a Ahmed’s (2006) idea o non-pe o ma i i y,
whe e e ms like inclusion a e used o hide hei opposi e, o he exclusion o ma -
ginalised g oups. This highligh s in e sec ional conce ns and a ocus on examining
he po en ially okenis ic na u e o e ms like s akeholde when i comes o ac ual-
ly empowe ing he g oups mos a ec ed by ins i u ional powe . We also consid-
3 Taylo , Vollme , Ra a , Benjamin
e ed E e Koso sky Sedgwick’s (2003) concep o pe i-pe o ma i i y, o unde s and
how he ways we alk abou hese p ac ices can shape wide unde s andings and
se he s age o u u e de elopmen s. This o e s impo an e lec ions on he dis-
cu si e alloca ion o pe o ma i e oles, and he sel - e e en ial elemen o go -
e nmen agendas in de ining who has access o policy. Ac oss hese heo ies, we
we e in e es ed in how he e m s akeholde migh be used in di e en ways o
p i ilege di e en g oups, o analyse wha powe dynamics i c ea es and o an-
swe ou co e ques ion o who is a s ake in UK ech policy.
LITERATURE ON STAKEHOLDERS
F eeman’s (1984) ounda ional ex on ‘managemen in u bulen imes’ si ua ed
s akeholde heo ies as eme ging om business and o ganisa ional con ex s, as a
coun e o ocusing solely on he p io i ies o sha eholde s and acknowledging a
wide social esponsibili y. The widesp ead ocus on s akeholde s could he e o e
be said o be less a ma e o ci ic du y and mo e oo ed in an o ganisa ional s a -
egy se ing business in e es s. The e m has also been shown o be p oblema ic
due o i s me cena y o igin (Sha s ein, 2016) o holding s akes in a be o ha ing a
inancial in es men in an issue, which p ecludes he wide ocus beyond sha e-
holding ha he heo ies a emp ed o p omo e. The e m s akeholde has also e-
cei ed c i icism due o i s conno a ions o s akeholde colonialism, in which “dis-
cou ses o co po a e ci izenship […] a e de ined by na ow business in e es s” and
s akeholde heo y “ ep esen s a o m o s akeholde colonialism ha se es o
egula ed he beha io o s akeholde s” (Banne jee, 2008). E en expansions in o
o ganisa ional e hics (Phillips, 2003) emain oo ed in ma ke -d i en mo i a ions
esponding o a cul u al and poli ical clima e ha began o demand and sc u inise
co po a e esponsibili y. Ac oss hese p oblema ic amings o s akeholde s as a
amewo k o o ganisa ional policy is a endency owa ds s aking a claim a he
han examining wha o who is a s ake. This o m o decision-making, hen, ac s
as a oil o dis ac ion, no necessa ily p omo ing he in e es s o anyone ou side
o exis ing posi ions o powe and in luence.
Phillips (2003) sugges s ha he powe in he e m s akeholde is due o i s con-
cep ual b ead h as he e m can be in e p e ed in a numbe o ways and holds di -
e en meanings o di e en people. Howe e , he emi o o ganisa ional e hics
places s akeholde ism as an op ional (i ideal) choice o a business, undamen ally
o i s own g ea e gain, a he han he policy and legal con ex s ha ca y a e-
sponsibili y o empowe di e en publics and p o ec he ulne able. An a ea
whe e hese conce ns ha e been demons a ed and discussed is public heal h.
4 In e ne Policy Re iew 14(3) | 2025
He e he b ead h o he e m has been exp essed as a se e ely limi ing ac o , en-
de ing s akeholde engagemen okenis ic (Sha s ein, 2016), wi h calls o g ea e
speci ici y in who exac ly is being coun ed as a s akeholde .
Wi hin ech policy, p e ious wo k on s akeholde s has also ocused on he go e -
nance o p i a e companies, pa icula ly hose labelled as ‘pla o ms’ (see, o ex-
ample, Go wa, 2019). Whe e s udies ha e examined he impac o s akeholde s on
legisla ion, egula ion and public policy, hey ha e ocused on global go e nance,
pa icula ly o In e ne p o ocols and in as uc u e (exis ing o en ou side o
s a es on p i a ely owned sys ems) (Belli, 2015; Eps ein e al., 2016; Ho mann,
2016; Lia opoulos, 2016), o pla o m egula ion and accoun abili y (exis ing wi h-
in he p i a e sec o in con lic wi h s a es) (Ba e and K eiss, 2019; Mannan and
Schneide , 2020) including ake news (Ma da and Milan, 2018), o cybe secu i y
and p i acy measu es (exis ing be ween s a es and wi h he p i a e sec o ) (e Sil a,
2013; Wol , 2016; Baue and an Ee en, 2009; Janssen e al., 2020), and o socie y-
wide p inciples and alues o he de elopmen o AI and ela ed eme ging ech-
nologies (especially ansna ional egula ion) (Ca h e al., 2018; Flo idi e al.,
2018; Hickman and Pe in, 2021; Smuha, 2021; Wilson, 2022). O he mo e speci ic
a eas ha e d awn on he applica ion o echnology o ields al eady in e es ed in
he ole o s akeholde s, such as AI in heal hca e (Sco e al., 2021) o online o -
ing (Bu kell and Regan, 2019). Niklas and Dencik (2021), in hei hema ic analysis
o he discou ses in s akeholde submissions o he EU Whi e Pape on AI, ound
ha . ‘ undamen al igh s’ was he mos p ominen o m o igh s (closely ollowed
by p i acy and human igh s), al hough e e ences o disc imina ion wi hin he
Whi e Pape i sel ocused on issues su ounding da a (bias, quali y, e c.) a he
han a deepe engagemen wi h he in e ac ions be ween uses o AI and igh s.
They also ound ha ‘social igh s’, while a g owing pa o wide discou se, was al-
mos absen om s akeholde discou ses. The indings o he s udy sugges how
con ibu ions by s akeholde s (which, as we ound in his s udy, we e domina ed by
co po a e en i ies) all back on dominan na a i es o legal minimum equi e-
men s a he han al e na i e ways o o ganising powe and echnology.
Howe e , s akeholde engagemen in such li e a u e is o en held as an answe o
co po a e and/o s a e powe . E en hough powe symme ies and dependencies
a e o en iden i ied ( o example, an Dijck e al., 2019), u he wo k is needed o
p oblema ise he no ma i e e ec s o in oking ‘s akeholde s’, and highligh ing he
powe asymme ies ha emain in dominan modes o s akeholde engagemen . In
much o he cu en egula o y landscape, mul is akeholde app oaches s ill allow
disp opo iona e in luence by p i a e companies o e egula ion, c ea ing loop-
5 Taylo , Vollme , Ra a , Benjamin
holes such as da a- a icking be ween ela ed co po a e en i ies (Kokas, 2023). In
his sense, s akeholde ism ac s as a code o sel - egula ion, which in u n co e-
sponds o in luence o e egula ion, essen ially allowing p i a e companies o
w i e hei own egula ion. A eas such as In e ne go e nance ha e iden i ied c i e-
ia o meaning ul engagemen (Malcolm, 2015). Bu while s akeholde ism is o en
p edica ed on he idea o an equal oo ing, his discu si e la ening also hides he
powe asymme ies ha emain, and he ole o he e m s akeholde in sus aining
such powe .
The e emains a ques ion o de ining he s akeholde , hough i is one ha we (a
leas pa ially) ejec . S akeholde can mean di e en hings in di e en con ex s,
and lacks a s able de ini ion. The e is g owing consensus a ound who he e m
s akeholde migh include; Go wa (2019), o example, o e s i ms, go e nmen ac-
o s and non-go e nmen al ac o s as a ipa i e sys em o o mal and in o mal
pla o m go e nance. Howe e , his o en alls back on hose who a e al eady en-
gaged in policy discussions o wield o he mechanisms o in luence echnology
p ac ices. I highligh s a sepa a ion be ween hose in ol ed ( o a ying deg ees) in
building policy and egula ion, and hose whom such policy is supposed o p o ec ,
hose subjec o he echnologies and policies. The claiming o a ‘s ake’ in policy
discussions - o he awa ding o such a s ake o ano he pa y - is i sel a ges u e
ha c ea es powe and oice o hose who a e o en al eady engaged in he con-
e sa ion. A key con ibu ion o his a icle is a shi in ocus om wha o who a
s akeholde is, bu on wha he e m s akeholde does, he di e en ways i is ap-
plied o di e en g oups and he di e en oles and powe s uc u es such a e m
cons i u es. This also leads o ou c i ique o he discou ses ha mobilise he e m
s akeholde h ough i s ocus on who holds a s ake, in o de o add ess he o en
obscu ed ma e o who is a s ake. We e u n o his ques ion in ou analysis and
discussion.
PERFORMING THE ROLE OF STAKEHOLDER
In o de o assess he e m s akeholde no as a s a e o being, wi h a s ake as a
hing o be ‘held’, bu ins ead as an ac i e ges u e ha cons i u es powe ela ions
in he way i is used, he heo e ical amewo k o his s udy is oo ed in quee
no ions o pe o ma i i y. The unde s anding o pe o ma i i y builds on ha de-
eloped by Judi h Bu le (1990; 2018) in ela ion o he social cons uc ion o gen-
de iden i y and applied by Ga ield Benjamin o echnology in ela ion o p i acy
(2020), us (2023), and he e minology su ounding da a (2021). This o m o
pe o ma i i y encapsula es he wo p ocesses o socially cons uc ing no ms and
6 In e ne Policy Re iew 14(3) | 2025
oles. On one hand, i is he way indi idual ac s o naming and speaking cons i u e
he iden i y o ole ha is being pe o med. This can be a pu pose ul o olun a y
ac o c ea ing iden i y by adop ing speci ic oles. On he o he , i is he social and
cul u al no ms ha shape, ein o ce and en o ce he oles ha people a e expec -
ed o pe o m. This can be an opp essi e se o s uc u es ha cons i u e and im-
pose powe asymme ies. In e ms o gende , hese p ocesses a e he way indi idu-
als cons i u e hei own iden i y h ough each ac o speech, d ess o manne ism,
as well as he se o social no ms ha a e simul aneously cons i u ed h ough
hese indi idual ac s and cons i u e he unequal oles ha indi iduals a e expec -
ed o pe o m. In e ms o he s akeholde , hen, we a e conce ned wi h ac s o la-
belling ha lay claim o a s ake in policy, o a oice in he discussions leading o
policy, and o powe and in luence o e he in e es s and alues embedded wi hin
such policy. While he de ini ions unde pinning ech policy a e s ill in lux, who is
in ol ed in de eloping a consensus on wha coun s as one echnology o ano he ,
and wha he pu pose o limi s o such echnologies should be, ca ies a imely im-
po ance in ela ion o no ma i e powe .
This pe o ma i e amewo k he e o e p o ides a use ul lens h ough which o ex-
amine he ole o he e m s akeholde in UK ech policy, wha he e m does o
cons uc ce ain powe ela ions (and no o he s). Each use o he e m helps con-
s i u e wha i means o be a s akeholde . This includes wha ha ole en ails,
wha powe o in luence i p o ides, and who is allowed o occupy ha ole. By ex-
ension, each ac o labelling o in oking s akeholde s con ibu es o claims o
who has a s ake in de eloping policy, whose oice coun s and whose in e es s a e
ep esen ed. Bu he wide ways in which making e e ence o s akeholde s is con-
s i u ed ac oss he gamu o policy documen s and p ocesses imposes no ms and
powe s uc u es ha pe pe ua e inequali ies in who coun s and whose oice is
ep esen ed. This shapes u he uses o he e m and no malises speci ic mean-
ings o he ole o s akeholde ha call in o ques ion who is eally being included.
This unde s anding o how “s akeholde ” is pe o ma i ely cons i u ed in ech poli-
cy can be ex ended by wo u he heo e ical de elopmen s. One o hese is i s in-
e se, wha Sa a Ahmed calls non-pe o ma i i y (2006). De eloping a c i ique o
ins i u ional inclusion ini ia i es, Ahmed ou lines how he ac o speaking a ole o
unc ion can be used o obscu e i s absence. Like di e si y and inclusion, he ole
o s akeholde can be used in ways ha signi y ep esen a ion o di e en in e -
es s wi hou making any changes o he powe s uc u es and p i ileges ha ocus
on he same iews ime and ime again. This includes he ague e e ences o
“wide s akeholde s” o “a ange o s akeholde s” o e en he mo e ob ious “o he
7 Taylo , Vollme , Ra a , Benjamin
s akeholde s” ha elega es such inclusion o an o he ed ca ego y o gene ic
ecogni ion, o en in he o m o consul a ion wi hou ue ep esen a ion.
The o he heo e ical posi ion ha in o ms ou discussion is pe i-pe o ma i i y as
ou lined by E e Koso sky Sedgwick (2003). This is he se o ac s abou ac s, he al-
lo- e e en ial p ac ices ha cons i u e no speech o oles hemsel es bu he con-
ex s in which hey a e p oduced. This pushes he collec i e no ms o pe o ma i i-
y u he by encou aging us o hink abou he ins abili y o powe s uc u es and
how e e ences o s akeholde s es ablish he con ex s in which hose s akeholde s
migh be ep esen ed. Pe i-pe o ma i i y comes in o play when he ole o s ake-
holde is pe o med on behal o he indi idual o g oup conce ned. This is a sig-
ni ican pa o he discou se o s akeholde engagemen in policy p ocesses as
s akeholde s a e o en e e ed o in gene ic o ca ego ical senses imposed on in-
di iduals o g oups, a he han being a ole o iden i y pu pose ully pe o med by
hose whose li es a e ‘a s ake’.
Wha his means o ou analysis is examining he di e en ways s akeholde s a e
in oked as oles, g oups o me hods o pe o ma i ely (o non-pe o ma i ely)
gene a ing he alues and in e es s behind ech policy. Wi h hese ools in hand,
we can me apho ically se he s age o he s akeholde o en e he spo ligh o
c i ique wi hin ou conside a ion o UK ech policy documen s.
METHODS
The s udy used a mixed me hod app oach, building quan i a i e and quali a i e
da a upon which o analyse he use o he e m s akeholde and he discu si e way
i cons i u es di e en o ms o powe and exclusion. Mixed me hods app oaches
ha e been used in echnology policy analysis o b idge gaps be ween exis ing
quan i a i e and quali a i e app oaches (Saheb and Saheb, 2024). We add o he
g owing body o li e a u e ga he ing echnology policy documen s h ough key
go e nmen al and in e go e nmen al sou ces (Co êa e al., 2023), u he honed
he e by he ocus on he pa icula egula o y con ex o UK public bodies. We ol-
lowed an induc i e app oach o documen and coding (Schi e al., 2021). Ou o-
cus on a speci ic heo e ical aming led o a mo e discou se analysis app oach
akin o wo k such as Niklas and Dencik (2021) h ough he lens o pe o ma i i y
(Benjamin, 2020; 2021) bu whe e hese men ioned s udies we e seeking key
hemes p esen in he documen s ou coding and analysis was mo e pu posi e in
ou ocus on “s akeholde ” as a e m.
8 In e ne Policy Re iew 14(3) | 2025
FIGURE 4: Dis ibu ion o sample documen s by sec o ca ego y and by echnology ype.
S akeholde
One hund ed and o y-one ou o he o al 194 documen s used he e m ‘s ake-
holde ’ (including plu als o hyphena ions such as mul i-s akeholde ). Figu es 5
and 6 show his o e ime. No able spikes a e 2019, which aligns wi h a ocus on
s a egy whe e engagemen wi h wide ex e nal ac o s may be mo e likely, as well
as an inc ease in 2022. The la e pe haps demons a es a mo e ecen ocus on
s akeholde engagemen , posi ioning he indings o his pape wi h inc easing im-
po ance as policy-make s seek o include - o o elude o including - mo e a ied
pe spec i es in how ech policy is implemen ed ac oss di e en sec o s. Ou ocus
is now on he 141 documen s ha men ioned s akeholde s. O hese, 70% made
e e ence o s akeholde s in pas engagemen , while 80% e e ed o hem as pa
o u u e plans ( o hemsel es o o he s). Sligh ly unde hal hese documen s
men ioned speci ic ypes o s akeholde s (46%), while he e we e also common
e e ences o “o he s akeholde s” (47%), “key” o “ ele an ” s akeholde s (41%) and
“di e se” o “a ange o ” s akeholde s (39%).The gene ic use o “s akeholde s” wi h-
15 Taylo , Vollme , Ra a , Benjamin
ou any quali ie was mo e dominan , being used by 81% o he documen s who
men ioned s akeholde s a all. To al numbe o uses places he gene ic use signi i-
can ly highe , a 703 compa ed o be ween 109 and 146 o mo e quali ied uses,
while pas and u u e we e mo e e en a 543 and 562 espec i ely. Figu e 7 shows
he ela i e p opo ion o di e en men ions ac oss empo al and use case ca e-
go ies.
FIGURE 5: Men ions o s akeholde compa ed o o al, by yea .
16 In e ne Policy Re iew 14(3) | 2025
FIGURE 6: A e age numbe men ions o s akeholde , by yea (o hose documen s men ioning he
e m).
Documen s wi h he highes use o he e m s akeholde we e mos ly go e nmen
depa men epo s de ailing esponses o consul a ions. These documen s a e im-
po an as hey show how he po en ially mo e inclusi e consul a ion p ocess
ansla es in o who is ac ually being ep esen ed as he policy is de eloped. They
we e, howe e , domina ed by gene ic uses o he e m s akeholde , no speci ying
who o e en wha ype o s akeholde was being e e enced. This speaks o a p ob-
lem o opaci y wi hin he policy-making p ocess, concealing in luence and no al-
ida ing ep esen a ion wi hin he consul a ion p ocess. I is no clea whe he
hese “s akeholde s” include he ull b ead h o submissions o only a ew inside
g oups al eady in ol ed in he discussion. This use o he e m s akeholde could
he e o e be conside ed non-pe o ma i e inso a as i emains una ibu ed. I
does no gi e he ole o s akeholde o a pa icula o ganisa ion, g oup o in e -
es , and he eby ails o acknowledge hose who ha e a s ake in he discussion. In-
s ead, ollowing he ins i u ional inclusion na a i es ha Ahmed c i iqued, his
use o he s akeholde in consul a ion becomes mo e a discu si e mobilisa ion o
he e m in suppo o exis ing policy agendas. I sugges s ha i is enough o lis-
en o “some” s akeholde s wi hou hinking abou “who” hose s akeholde s ac u-
17 Taylo , Vollme , Ra a , Benjamin
ally a e.
FIGURE 7: Two cha s showing pe cen age usage o he e m s akeholde , i s ly by coded ype,
secondly by empo al ame.
The highes use o speci ic s akeholde s was in ela ion o business opics. This
demons a es he his o ical use o he s akeholde as ha ing a inancial s ake,
alling back on he de ini ions o colonial o co po a e s akes, owne ship and in lu-
ence discussed abo e (Sha s ein, 2016; Banne jee, 2008). The ocus on indus ial
s a egy and g ow h highligh s a poli ical agenda d i en by speci ic economic p in-
ciples, suppo ed by ou obse a ions on documen s such as he Compe i ion and
Ma ke s Au ho i y epo s on digi al ad e ising on online pla o ms, which p io i-
ised he e e ences o he “indus y” s akeholde s. This use o speci ic s akehold-
e s gi es a s ake, gi es a oice, o hose al eady in con ol o such sys ems a he
han he use s and especially ma ginalised g oups who would be a ec ed by hei
decisions.
O he opics wi h a high use o he e m s akeholde o e all we e s a egy, which
is app op ia e gi en he consul a i e p ocess al eady men ioned, as well as in o -
ma ion secu i y opics, which builds on mul i-s akeholde ism as a key pa o op-
ics such as in e ne go e nance, na ional secu i y and suppo ing smalle o ganisa-
ions in hei cybe secu i y p ac ices. O he as majo i y who did e e o s ake-
holde s in some way, mos e e ed o hem many imes o e . The ew o ganisa-
ions who ne e made any e e ence o s akeholde s we e hose wi h only one o
wo epo s in he da a se . As ou sampling o policy documen s was conduc ed
p io o iden i ying hei use o he e m s akeholde , his is unlikely o be a sam-
pling a e ac . Ins ead, his inding is mo e likely due o hese examples being
mo e speci ic applica ions wi hin o he policy sec o s, whe e he wo d s akeholde
may no be used in a ou o naming speci ic in e es ed and in ol ed pa ies. This
18 In e ne Policy Re iew 14(3) | 2025
demons a es he po en ial o a mo e accoun able app oach owa ds engaging
wi h s akeholde s by naming di ec ly hose who a e in ol ed in decision-making
and hose who a e a ec ed.
Some documen s, such as hose by he NHS, NDG and NAO we e almos i no ex-
clusi ely e e ing o pas engagemen , which also (wi h a ied usage) demon-
s a es a ce ain le el o accoun abili y in including di e en pe spec i es in he
policy p ocess. Bu many o he s, including many o hose by he ICO, CDEI, O Com,
BEIS, DHSC, and HM T easu y, lean much mo e hea ily owa ds u u e impe a i es
(in he case o ICO in pa icula , his was o en as an ins uc ion o o he o ganisa-
ions a he han hei own plans). This shows a spli in e ms o hose making
g and s a egies o o he s o implemen o o wo k ou de ails la e , e sus he
mo e ope a ional side o building om exis ing ela ionships. Take he di e ence
be ween DHSC and he NHS, o example: he go e nmen depa men ocused on
o e a ching s a egy and expec a ion ha s akeholde engagemen would ollow,
while he heal h se ice al eady embedded wi h p o essionals, pa ien s and suppli-
e s b ings ha engagemen in o i s policy documen s.
These indings show ha he use o s akeholde ends o pe o ma i ely en ench
he exis ing powe o “indus y s akeholde s” o nameless bu clea ly al eady en-
gaged and empowe ed “key s akeholde s”. Meanwhile, hey also cons uc a alse
sense o inclusion h ough he non-pe o ma i e use o gene ic o “o he s ake-
holde s”. This c ea es signi ican isk o a eil o accoun abili y, and aises signi i-
can ques ions o e es ablished p ocesses such as consul a ion. When i is unclea
who is in luencing policy, whose oices and in e es s a e being ep esen ed, hen
he indica o s om speci ic uses sugges ha he s akeholde becomes a oil o
ampli ying his o ical powe and p i ilege, o en on poli ical and/o economic lines,
and in doing so excludes he needs o hose mos a ec ed by echnologies who al-
eady su e a lack o agency in how da a, AI, pla o ms and o he a eas a e used o
shape hei li es.
Ci a ions
How do hese uses o he e m s akeholde compa e o he oices ac ually (and
iden i iably) ep esen ed in policy documen s? We e u n now o he ull da a se
o 194 documen s o conside which sou ces we e being ci ed wi hin he docu-
men s. Figu e 8 p o ides an o e iew o he o al ci a ion sou ces ep esen ed in
he da a se o e ime. Hal (50.4%) o all ci a ions ac oss all epo s assessed we e
om go e nmen depa men s, while almos wo hi ds (125/194) had 50% o
mo e sou ces om o he go e nmen epo s, showing a ci cula i y and sel - epe i-
19 Taylo , Vollme , Ra a , Benjamin
ion ac oss go e nmen depa men s and policies. Only 13 documen s had no ci a-
ions o go e nmen sou ces, hough all bu wo o hese had no ci a ions a all.
These 11 documen s ha had no ci a ions included h ee ha we e based on di-
ec su eys, while he emainde we e s a egies and s anda ds in a eas like De-
ence. The mos ci ed sou ce ou side o go e nmen was indus y, a 15% o all ci-
a ions, ollowed by academia wi h 13%, media wi h 8%, ci il socie y wi h 7% and
go e nmen (in e na ional) wi h 6%.
FIGURE 8: Numbe o ci a ions e e enced in sample documen s by sou ce ype, by yea .
The dominance o go e nmen sou ces shows a consis en and common agenda
h ough au ho i a i e sou ces. I ollows ha la ge o s a egic documen s would
go on o shape many u u e policy documen s, o en in mo e speci ic a eas o im-
plemen a ion. This acks wi h how ech policy has sp ead ac oss UK go e nmen
depa men s as mo e a eas ha e had o g apple wi h he impac o da a, AI, pla -
o ms and he like on hei pa icula emi . Howe e , his insula sou cing o iews
ca ies a high isk o poli ical single-mindedness and a closed agenda. Once pa -
icula s a egies a e pu in place, hey echo ac oss di e en policy a eas, which
po en ially o e ides he needs o speci ic con ex s – speci ic s akeholde s and
speci ic people o g oups whose li es a e ‘a s ake’ – no o men ion he wide ex-
pe ise ha could be b ough in o in o m hese mo e speci ic a eas. While i is
again pe haps an e ec o s anda d policy p ac ices, he homogenisa ion and se -
ing in s one o policy agendas sugges s he exclusion o al e na i e oices, pa ic-
ula ly hose mos ele an o speci ic implemen a ions o policy s a egy.
20 In e ne Policy Re iew 14(3) | 2025
We ound ha highe le els o go e nmen sou ces ended o co espond o lowe
inclusion o ex e nal sou ces and less di e si y be ween sou ces (based on s an-
da d de ia ion o ela i e ci a ion coun s gene ally being signi ican ly highe o
such documen s and quali a i e obse a ions o he sp ead ac oss o he ca e-
go ies). This leads o g ea e une enness be ween oices, compa ed o documen s
showing mo e o any non-go e nmen sou ce which ended o indica e a mo e di-
e se and e en a ay o ci a ions ac oss di e en sec o s. The e o e, documen s e-
lian on go e nmen sou ces end o escala e he insula i y and ci cula i y ei he
in e nally o wi h speci ic ‘inside ’ g oups. The dominance o in e nal ci a ion sug-
ges s ha es ablished go e nmen na a i es become en enched ac oss depa -
men s and homogenise policy agendas and p io i ies, which could be conce ning
when speci ic a eas like heal h o educa ion equi e mo e di ec ed engagemen
wi h sec o -speci ic knowledge. Howe e , i was eassu ing in his sense ha
heal h, and social o e hical issues, we e among he a eas ha mos highligh ed
wo k by ci il socie y and academia. Se en such epo s elied mo e on academia
han o he sou ces combined. Meanwhile, he epo s by he Child en’s Commis-
sione we e among he mos elian on ci il socie y, and in e es ingly hese didn’
men ion s akeholde s, despi e he Child en’s Commissione ha ing a end owa ds
highe le els o engagemen wi h a ec ed g oups (speci ically, child en and pa -
en s).
Fou epo s we e mo e elian on indus y han o he sou ces combined. The e-
po s ha had a highe han 1 s anda d de ia ion abo e a e age o indus y
sou ces we e pe haps unsu p isingly wi hin he a eas o business, inance and cy-
be secu i y. Mo e su p ising pe haps was ha epo s by de ol ed go e nmen s
we e also in his g oup, which shows he ocus o egional in e es in echnology
and/o he a eas whe e de ol ed go e nmen s a e able o claim mo e agency o e
hei engagemen wi h echnologies. Media epo ing was unsu p isingly highes
conce ning online habi s and pla o ms. I we ake in o accoun he p i a e owne -
ship o media ou le s and he e o e combine hem wi h indus y, howe e , hen we
see his g oup domina ing he in luence o e : he egula ion o online pla o ms
(in which hey ha e signi ican sel -in e es ); business s a egy; and, pe haps mos
ala mingly, he UK Digi al S a egy 2017.
The epo wi h he mos e enly dis ibu ed sou ces ac oss he ca ego ies was he
ICO’s ‘Big da a, AI, machine lea ning and da a p o ec ion’. This is in e es ing as he
ICO gene ally elies on a la ge numbe o sel -ci a ions, poin ing eade s o i s
weal h o guides and epo s on he minu iae o da a p o ec ion obliga ions.
Agains his gene ally na ow ocus a e epo s such as his one which show a
21 Taylo , Vollme , Ra a , Benjamin
b oade engagemen wi h, o example, academia and ci il socie y in es ablishing
he g oundwo k o u u e policy wo k. This sugges s ha he e a e key momen s,
key documen s whe e b oade in luence is possible and encou aged.
The UK Digi al S a egy 2017 and 2022 show an in e es ing di e gence. Whe e he
2017 s a egy was a ied in i s sou ces, by 2022 his had na owed signi ican ly o
o he go e nmen documen s. This could pa ially be explained by he inc ease in
go e nmen wo k on he a ea du ing ha ime, so he e was simply mo e o d aw
on, o ha he 2022 s a egy was mo e o an upda e han new policy. Bu he pe i-
od also saw u he wo k by academia, ci il socie y and o he s, so i s ill ep esen s
a shi in ocus. Simila shi s owa ds cen alised go e nmen sou ces o e ime
can be seen in he di e ence be ween he Na ional AI S a egy and he s a egy’s
Ac ion Plan he ollowing yea ; while bo h we e domina ed by go e nmen
sou ces, his was e en mo e p onounced (58% o 79%) in he ac ion plan. The
same shi was also seen in he di e ence be ween in e im and inal epo s, such
as he CMA’s Online Pla o ms and Digi al Ad e ising Ma ke s udy. The in e im
epo was mo e di e se, wi h minimal de ia ion seen only in indus y (abo e a e -
age) and ci il socie y (unde ), whe eas he inal epo showed he end owa ds
go e nmen sou ces, in his case alongside indus y.
This, in he con ex o he o e all dominance o go e nmen sou ces, sugges s ha
he e a e key momen s – o en mo e unde lying esea ch epo s – whe e ex e nal
pa ies can gain in luence o make al e na i e oices hea d. Then, once hese a e
se in s one, he engagemen na ows owa ds a mo e ocused go e nmen na a-
i e. This aises conce ns su ounding he me hod o engagemen wi h s akehold-
e s, pa icula ly when consul a ions lis all submissions bu do no di ec ly ma ch
hese in a anspa en way o ac ual go e nmen policy.
DISCUSSION
This is a use ul poin o e isi he meaning o he e m s akeholde wi hin UK ech
policy. Ou indings show ha a consis en de ini ion emains impossible o pin
down. Howe e , ou analysis wi hin a pe o ma i e amewo k is ins uc i e in say-
ing, i no wha a s akeholde is hen wha he e m s akeholde does o gi e a
s ake o ce ain ypes o ac o s and no o he s. We he e o e o e he ollowing di-
ision o he ways s akeholde is used:
Pe o ma i e uses o he e m s akeholde a e hose ha gi e a s ake, gi e oice, o
pa icula ac o s o g oups. This is seen in speci ying speci ic pa ies, such as “in-
dus y s akeholde s” o “go e nmen s akeholde s”, which makes up almos a qua -
22 In e ne Policy Re iew 14(3) | 2025
e o occu ences. This use cons i u es hose g oups as ha ing a s ake, o en align-
ing wi h inancial incen i es o businesses being egula ed o public sec o bodies
whose expendi u e may be a ec ed. When his was obse ed in s a egy-se ing
epo s, his sugges s a cen alisa ion o agendas and dep io i isa ion o he
g oups mos ma ginalised by he use o he echnology. This is seen in ou indings
wi h he highes ins ances o speci ic named s akeholde s being used in business
and by o ganisa ions like he Compe i ion and Ma ke s Au ho i y, which aises
ques ions abou in luence and sugges s an unwillingness o egula o s o ake a
s ong s ance in en o cing laws a ound da a p o ec ion and algo i hmic manipula-
ion, o example. Some imes his use may deno e s a us wi hou speci ici y, such
as “key s akeholde s”, implying an opaque lis o s akeholde s able o wield powe .
These uses o en align wi h pas engagemen wi hin he policy p ocess, designa -
ing a ole o inside and g an ing in luence wi hou anspa ency o accoun abili y.
Non-pe o ma i e uses a e hose ha deny a s ake while making gene ally ague
e e ence, pe o ming a legi imising unc ion. I o en aligns wi h u u e plans o
expec a ions, a gene alised aim a he han a speci ic commi men o conc e e ac-
i i ies, and wi h desc ip ions such as “wide s akeholde s” o “di e se s akehold-
e s”. While his language could be used o lea e open he ques ion o possible in-
ol emen , he endency o pas engagemen o na ow down on o expec ed
g oups al eady ea u ing hea ily in ci a ional p ac ices sugges s i is no pe o m-
ing ha unc ion. We ound ha while he men ion o an unspeci ied “ ange”
(“wide ”, “di e se”) s akeholde s was a less han ⅛ o cases, i aligned hea ily wi h
u u e ense. The non-pe o ma i e s akeholde is he e o e a nonspeci ic s a e-
men o noncommi al in en . A s onge use ha pe o ma i ely cons i u ed inclu-
sion would speci y which g oups will be included in u u e policy decisions, e en i
space was le open o engagemen wi h as ye unde e mined g oups (indeed, his
dual app oach would be encou aged o ou line a clea and anspa en plan o in-
clusion while also ecognising he need o keep adap ing and expanding whose
oices a e hea d). The non-pe o ma i e “ ange o ” s akeholde s is ins ead mo e an
embodimen o shallow inclusion ini ia i es ha eign openness while es ic ing
he ac ual in luence o hose o en mos a ec ed. In ech policy ha ends o in-
clude use s, ulne able g oups such as hose om quee and ans* communi ies,
o hose ma ginalised by ace, class, educa ion and/o gende . The lack o speci ici-
y elides accoun abili y o he exclusion o such g oups, while he use o he e m
s akeholde in bo h pe o ma i e and non-pe o ma i e ways c ea es he discu -
si e illusion o equali y.
Pe i-pe o ma i i y in he use o he s akeholde is he hi d-pe son alloca ion o
23 Taylo , Vollme , Ra a , Benjamin
he ole o o he s. The combina ion o pe o ma i e and non-pe o ma i e oles
enac ed in he policy documen s we s udied shows a se o speech ac s ha en-
able some g oups o pe o ma i ely cons i u e hei own ole as s akeholde , o
claim hei own s ake, while non-pe o ma i ely denying he same s akes o o h-
e s. The o me is o en hose al eady holding he esou ces, in luence and powe
o make hei oice known h ough in i a ions o ound ables, clea pa hways o
ci a ion, named iden i ica ion as a “key s akeholde ” o au oma ic inclusion in a
discussion. The la e , howe e , is o en hose who a e a s ake in such discussions,
and who ins ead ely on he ep esen a ion o hei in e es s by hi d pa ies, o en
ci il socie y o academic g oups who a e hemsel es less ep esen ed in ci a ional
p ac ices han indus y ac o s. O he imes, he absence o speci ici y may cons i-
u e in luence, such as epo au ho s poin ing o unnamed s akeholde s whose
con ibu ions align wi h he p ede e mined agenda, no able in ou indings in he
che ypicking ha occu s in go e nmen esponses o consul a ions o in he oll-
ou o policies om he Cen al Digi al and Da a O ice (pa o he cen alised
Cabine O ice) o o he sec o s. This use o he s akeholde he e o e also high-
ligh s he sel - e e en ial na u e o go e nmen policy. While cen al s a egy ol-
lowed by implemen a ion in speci ic policy a eas is o be expec ed, he common
lack o engagemen wi h sec o -speci ic a ec ed g oups is conce ning. Simila ly,
he dec ease in ex e nal pe spec i es in cen al s a egy o e ime sugges s a clos-
ing o o who can be deemed a s akeholde , and an en enching o he s akes
along poli ical and inancial lines. This isk he e is no me ely ha policy agendas
become ixed bu ha who has a say becomes pe i-pe o ma i ely es ic ed and
i e a ed.
Figu e 9 shows an indica i e dis ibu ion o hese ypes o s akeholde s acco ding
o he ca ego ies eme ging om he da a se as well as ou analysis o who is im-
plied bu no men ioned in he discou se o s akeholde inclusion. I is by no
means a comple e mapping, bu ac s as a guide o he landscape o who is a o d-
ed a s ake in ech policy in di e en se ings and who is a s ake in he esul s o
he policy p ocess. This diag am shows he di e gence e en wi hin go e nmen as
well as he a ied scope o o he sec o ac o s. Those wi h mo e powe o be able
o pe o m he ole o s akeholde a e able o in luence policy in a mo e di ec
way. Those able o pe o ma i ely hold a s ake a e o en hose mos likely o bene-
i om a policy, while hose non-pe o ma i ely denied i a e o en hose a s ake
in he po en ial ha ms. This asymme ic dis ibu ion o s akes aises ques ions
abou he alue, accoun abili y and legi imacy o exis ing policy p ocesses. This is
ue in policy mo e widely, bu he pa icula con es a ions o e ms, alues and
aims wi hin ech policy make his an all he mo e pe inen a ea o conce n.
24 In e ne Policy Re iew 14(3) | 2025
Van Dijck, J., Niebo g, D., & Poell, T. (2019). Re aming pla o m powe . In e ne Policy Re iew, 8(2). h
ps://doi.o g/10.14763/2019.2.1414
Wilson, C. (2022). Public engagemen and AI: A alues analysis o na ional s a egies. Go e nmen
In o ma ion Qua e ly, 39, 3–10. h ps://doi.o g/10.1016/j.giq.2021.101652
Wol , J. (2016). Wha we alk abou when we alk abou cybe secu i y: Secu i y in in e ne
go e nance deba es. In e ne Policy Re iew, 5(3). h ps://doi.o g/10.14763/2016.3.430
31 Taylo , Vollme , Ra a , Benjamin
