CoARA Wo king G oup on ‘E hics and Resea ch In eg i y
Policy in Responsible Resea ch Assessmen o Da a and
A i icial In elligence (ERIP)’
P o . D . Pe ihan Eli EKMEKCI
•Wo ks eam 1: The c ea ion o an ou each p og am o shape policy, os e ins i u ional and socie al change, and c ea e awa eness, p omo ing a cul u e o
esponsible da a and AI esea ch.
•oThe ERIP Handbook on in eg i y, da a, and AI in esea ch e alua ion o o e a comp ehensi e guide o na iga ing e hical complexi ies and le e aging digi al
ad ancemen s in con empo a y esea ch assessmen p ac ices.
•oA Se ies o ERIP Webina s
•oS akeholde Engagemen Wo kshops
•oSocial Media P esence
••Wo ks eam 2: The de elopmen o policy, me hods, and guidelines o ensu e he esea ch e hics and in eg i y o scien i ic me hods and ou pu s wi h he
ad ancing use o da a and applica ion o AI me hods and o assess digi al con ibu ions o science/knowledge in esea ch p og ams and assessmen p ocedu es.
•oA Policy F amewo k o E hical Resea ch Assessmen in he Digi al Age
•oA F amewo k o assessing and e alua ing Digi al Con ibu ions .
•oMe hodological Guidelines o Assessing Digi al Con ibu ions and E hical Resea ch Assessmen in he digi al age
• •Wo ks eam 3: The de elopmen o a comp ehensi e oolki o he implemen a ion o bes p ac ices in esea ch assessmen o da a and AI.
•oA oolki o e alua ion o Digi al Con ibu ions o Science
•oA Digi al Toolki o Responsible Resea ch Assessmen
•oTechnology-based Tools o Bes P ac ices o Da a and AI in Resea ch Assessmen
Objec i es o ERIP
• o de elop e hics and esea ch in eg i y guidance o esea ch assessmen in ol ing da a and AI;
• o p omo e esponsible da a and AI p ac ices by ad oca ing o esponsible assessmen p ac ices
ha p io i ize in eg i y, esponsibili y, and socie al bene i , while mi iga ing he po en ial isks and
ha ms associa ed wi h da a and AI-d i en esea ch;
• o os e s akeholde engagemen h ough acili a ed dialogue and collabo a ion among
esea che s, policymake s, unding agencies, and o he ac o s o ensu e di e se pe spec i es in
he de elopmen and implemen a ion o e hical esea ch assessmen policies;
• o p o ide a oolki o educa ional esou ces o aise awa eness and help implemen bes p ac ices
in esea ch assessmen o da a and AI; and
• o engage in ou each o shape policy and ins i u ional change, p omo ing a cul u e o esponsible
da a and AI esea ch.
De eloping me hods and ools
o ensu e he esea ch e hics and
in eg i y o scien i ic ou pu s wi h
he ad ancing use o da a and he
impac o AI
o e alua e digi al con ibu ions o
science/knowledge in esea ch
p og am and assessmen
p ocedu es; and
inno a i e me hodologies o
employing da a ecosys ems and AI
models o esea ch assessmen in
digi al en i onmen s.
Guiding p inciples o help us na iga e he comp
lexi ies o esea ch assessmen in he digi al ag
e.
Embedding E hics in o AI P ac ices o
Resea ch Assessmen
The Modula Con igu a o Suppo ed by he ENVISAGE
P inciples and PILOT P inciples
Open Science Fai Session
E hics Fi s : Re hinking Resea ch Assessmen in he Age o AI and Open Science
CERN, Gene a, Swi ze land
17 Sep embe 2025
F ancis P. C awley
COARA-ERIP
CODATA IDPC
The Why We Need an E hical O e haul
Ou da ed Me ics, New Reali ies
T adi ional me ics (JIF, ci a ions) ail o assess mode n esea ch: da ase s, so wa e, in e disciplina y
wo k, and socie al impac .
The AI Re olu ion in oduces new challenges:
•Asymme y: AI-empowe ed applican s s. assessmen p ocesses lacking ools o e alua e AI-
gene a ed con en .
•New a i ac s: How o assess AI models, gene a ed con en , and complex digi al ou pu s?
•E hical isks: Ampli ica ion o bias, lack o anspa ency, and h ea s o esea ch in eg i y.
The Co e Con lic : The gap be ween apid echnological ad ancemen and slow, igid assessmen
amewo ks isks unde mining science and highe educa ion.
OSF Gene a CERN, 17 Sep 2025 Embedding E hics in o AI P ac ices o RA Slide 6
The Modula Assessmen Con igu a o (MAC):
A Vision o E hical, Adap i e Resea ch Assessmen
F om Rigid Me ics o Responsible Con igu a ion
The MAC is a p oposed digi al, in e ac i e ool designed o ope a ionalize high-le el e hical p inciples
in o p ac ical, cus omizable esea ch assessmen (RA) p ocesses. I mo es beyond one-size- i s-all
me ics o allow o con ex -sensi i e, alue-based e alua ion ac oss a ange o esea ch con ibu ions
and a i ac s.
Key Mo i a ions:
•The P oblem: T adi ional me ics (JIF, ci a ion coun s) a e inadequa e o assessing mode n,
collabo a i e, and AI-enabled esea ch ou pu s like da ase s, so wa e, and AI models.
•The AI Challenge: C ea es an asymme y be ween AI-empowe ed applican s and assessmen
p ocesses ha lack he ools o e alua e new esea ch a i ac s e hically and anspa en ly.
•The Need o Flexibili y: Di e en disciplines, ins i u ions, and missions equi e di e en assessmen
c i e ia. The MAC p o ides he needed con igu abili y.
OSF Gene a CERN, 17 Sep 2025 Embedding E hics in o AI P ac ices o RA Slide 7
The Modula Assessmen Con igu a o (MAC):
A Vision o E hical, Adap i e Resea ch Assessmen
F om Rigid Me ics o Responsible Con igu a ion
The MAC is a p oposed digi al, in e ac i e ool designed o ope a ionalize high-le el e hical p inciples
in o p ac ical, cus omizable esea ch assessmen (RA) p ocesses. I mo es beyond one-size- i s-all
me ics o allow o con ex -sensi i e, alue-based e alua ion ac oss a ange o esea ch con ibu ions
and a i ac s.
Key Mo i a ions:
•The P oblem: T adi ional me ics (JIF, ci a ion coun s) a e inadequa e o assessing mode n,
collabo a i e, and AI-enabled esea ch ou pu s like da ase s, so wa e, and AI models.
•The AI Challenge: C ea es an asymme y be ween AI-empowe ed applican s and assessmen
p ocesses ha lack he ools o e alua e new esea ch a i ac s e hically and anspa en ly.
•The Need o Flexibili y: Di e en disciplines, ins i u ions, and missions equi e di e en assessmen
c i e ia. The MAC p o ides he needed con igu abili y.
OSF Gene a CERN, 17 Sep 2025 Embedding E hics in o AI P ac ices o RA Slide 8
The E hical Compass
ENVISAGE P inciples o DRR
ENVISAGE p o ides he cons i u i e e hical ‘why’ o using AI in DRR, ensu ing echnology se es
humani y. C i ical p inciples o c isis con ex s:
•So e eign y (S): Respec s communi y & na ional igh s o e hei da a and isks.
•In eg i y (I): Ensu es hones y and eliabili y o AI-d i en isk in o ma ion.
•Access (A): P omo es equi able access o AI ools and bene i s o all communi ies.
•Engagemen (E): Manda es in ol ing local s akeholde s in AI design and deploymen .
•Goal: P e en s digi al colonialism, builds us , and ensu es AI solu ions a e cul u ally and
con ex ually app op ia e.
‘To en isage is no me ely o imagine bu o an icipa e wi h e hical pu pose and o ien ac ion owa d jus and humane ou comes.’
OSF Gene a CERN, 17 Sep 2025 Embedding E hics in o AI P ac ices o RA Slide 9
E hics-Based Tools o Responsible Assessmen
Based on CoARA-ERIP’s wo ks eams:
❑How can we ensu e anspa ency and accoun abili y when assessing esea ch ou pu s gene a ed o
assis ed by AI? Should AI-gene a ed con en be e alua ed di e en ly om human-au ho ed wo k, o by he
same e hical s anda ds?
❑A e ou cu en e alua ion sys ems equipped o ai ly ecognize he alue o in e disciplina y and
ansdisciplina y esea ch? Wha biases o blind spo s migh exis in adi ional me ics when applied o
collabo a i e o c oss-sec o al wo k?
❑Wha would an e hically inclusi e esea ch assessmen amewo k look like in p ac ice?
How can we balance quali a i e and quan i a i e indica o s o e lec in eg i y, equi y, and socie al
ele ance?
CoARA-ERIP is de eloping amewo ks ha espond o hese challenges. These include:
▪E hical c i e ia o e alua ing AI-gene a ed esea ch ou pu s, ensu ing anspa ency and accoun abili y.
▪Quali a i e indica o s ha e lec in eg i y, equi y, and socie al ele ance.
▪Tools o assessing in e disciplina y and ansdisciplina y con ibu ions, which a e o en o e looked in
adi ional models.
These amewo ks a e no abs ac ideals; hey a e p ac ical ins umen s ha ins i u ions, unde s, and
e alua o s can adop o build mo e jus and e lec i e assessmen sys ems.
E hics-Based Tools o Responsible Assessmen
A conc e e example:
Imagine a uni e si y e alua ing candida es o a esea ch posi ion. Ins ead o elying solely on
publica ion coun s o jou nal ankings, i could apply CoARA-ERIP’s e hical c i e ia o assess he
socie al impac , collabo a i e na u e, and me hodological in eg i y o each candida e’s wo k.
F om a bioe hical s andpoin , esponsible esea ch assessmen mus be guided by co e p inciples:
❑Au onomy: Respec ing di e se me hodologies and epis emologies.
❑Jus ice: Ensu ing ai e alua ion ac oss disciplines and geog aphies.
❑Bene icence and non-male icence: A oiding ha m by ecognizing and aluing con ibu ions ha may
be in isible o con en ional me ics.
Ins i u ions ha e a mo al esponsibili y o os e a cul u e o e hical e alua ion, one ha suppo s
inno a ion wi hou comp omising in eg i y.
A Bioe hical Pe spec i e
Respec o di e se me hodologies, epis emologies, and esea ch cul u es.
P omo es ecogni ion o non- adi ional ou pu s, including AI-assis ed and collabo a i e esea ch, ensu ing
esea che s a e no penalized o choosing inno a i e o inclusi e app oaches.
A e we allowing esea che s he eedom o inno a e, o a e we cons aining hem wi h ou da ed me ics?
Au onomy
Jus ice
Fai e alua ion ac oss disciplines, geog aphies, and ins i u ional con ex s. Ad oca es o inclusi e
assessmen sys ems ha ecognize in e disciplina y and ansdisciplina y con ibu ions and add ess
sys emic biases in global esea ch e alua ion.
- Only 12% o exis ing amewo ks adequa ely assess in e disciplina y wo k, e ealing a jus ice gap in
cu en p ac ices.
How can we edesign e alua ion sys ems o ensu e equi y ac oss di e se esea ch landscapes?
A oiding ha m by ecognizing and aluing con ibu ions ha may be in isible o con en ional me ics.
Encou ages he use o quali a i e indica o s ha e lec socie al impac , e hical in eg i y, and collabo a i e
alue, especially in AI-gene a ed and da a-in ensi e esea ch.
AI-gene a ed con en poses isks o misin o ma ion and bias. E hical e alua ion amewo ks mus ensu e
hese ou pu s a e assessed wi h anspa ency and accoun abili y.
A e we unin en ionally ha ming scien i ic p og ess by igno ing aluable con ibu ions ha all ou side
adi ional me ics?
Bene icence and Non-Male icence
Ins i u ions ha e a mo al and s a egic esponsibili y o aise a cul u e o e hical e alua ion, one ha suppo s
inno a ion wi hou comp omising in eg i y.
This includes:
•Adop ing CoARA-ERIP’s ools and amewo ks;
•T aining e alua o s in e hical and inclusi e assessmen ;
•Engaging esea che s in co-c ea ing e alua ion c i e ia.
Wha s eps can you ins i u ion ake oday o embed e hics in o i s esea ch assessmen p ac ices?
Conclusion –E hics Fi s
E hics is no a ba ie o p og ess; i is he ounda ion o us wo hy science.
As we e hink esea ch assessmen in he age o AI and open science, we mus place e hics a he cen e
o ou amewo ks, policies, and p ac ices.
I in i e all o you— esea che s, unde s, ins i u ions- o engage wi h he wo k o CoARA-ERIP and help
build sys ems ha e lec he alues we wan science o uphold.
To e alua e e hically is o ecognize he ue alue o science—no jus o academia, bu o socie y.
By placing e hics a he hea o esea ch assessmen ,
CoARA-ERIP is no jus eshaping how we e alua e
science, i is ede ining how science se es socie y
wi h in eg i y, inclusi eness, and pu pose.
Call o Ins i u ional Ac ion
E hics Fi s :
Re hinking
Resea ch
Assessmen
15- 17 SE P T E M BE R
Au ho s:
M CRAWLEY, F ancis P. (CoARA E hics and Resea ch In eg i y Policy Wo king G oup on Responsible Resea ch Assessmen in
Da a and A i icial In elligence (CoARA-ERIP));
P o . EKMEKCI, Pe ihan Eli , (Associa e P o esso , Depu y Dean, and Head o he His o y o Medicine and E hics Depa men ,
School o Medicine, TOBB ETU)
P o . DE SOUSA FREITAS, Ma a (Di ec o o he Ins i u e o Bioe hics o he Uni e sidade Ca ólica Po uguesa (IB-UCP));
