Anika Mülle -Ka abil | Ma ie Ala i
OSC 2025, Hambu g
Open Science Capaci y Building in imes o AI:
Finding Solu ions wi h he GATE
Hambu g Oc obe 2025
Suppo ed by:
We a e in a solu ion session.
So wha a e ou challenges/p oblems?
Pu pose o his session
#challenge 1:
OS is con inuously e ol ing.
OS
OS
OS
Access
RCR, RE
AI, digi al
Secu i y, access e c.
AI, digi al
Access
Access
RCR, RE
RCR, RE
Access.
OS
Access
RCR, RE
AI, digi al
Secu i y, access e c.
?
Challenges
OS
#challenge 1:
OS is con inuously e ol ing.
OS
OS
OS
Challenges
OS
#challenge 2:
Ad ancing OS equi es ac ion.
Especially in imes o AI.
OS
GATE suppo s con inuous knowledge
exchange on OS p ac ices & guiding hough s.
GATE helps - #challenge 1
UNDERSTANDING OS
OS
OS
OS
#challenge 1:
OS is con inuously e ol ing.
#challenge 2:
Ad ancing OS equi es ac ion.
Especially in imes o AI.
GATE
Communi y
GATE
Capaci y
GATE
Da a
DOING OS
UNDERSTANDING OS
How i wo ks - S ands o he GATE
GATE
Resea ch
GATE
Da a
GATE
Communi y
GATE
Capaci y
UNDERSTANDING OS
How i wo ks
How i wo ks
GATE Se ice
GATE
Communi y
GATE
Capaci y
UNDERSTANDING OS
Knowledge
c ea o s
Resea ch
communi ies
Da a
in as uc u es
RFOs
Policy
Resea che s
How i wo ks
GATE Se ice
GATE
Communi y
GATE
Capaci y
GATE
Communi y
UNDERSTANDING OS
Quick session walk- h ough
helpe 1: Pe sona
a he g oup ables
In you g oup, choose a
pe sona
a ge g oup you a e c ea ing a solu ion o
WHO IS YOUR ACTION FOR?
2
Quick session walk- h ough
helpe 1: Pe sona
a he g oup ables
In you g oup, choose a
pe sona
a ge g oup you a e c ea ing a solu ion o
WHO IS YOUR ACTION FOR?
Senio esea che
ECR
Funde
Open Access O ice
Da a S ewa d
Lib a ian
2
D . Tomasz, 56, Senio Resea che in [e.g. Medicine]
Challenges
Needs
“Open Science? I’ e always published in
high-quali y op jou nals - I doub he e
a e high-quali y open-access jou nals?”
“AI ools look in e es ing, bu I need
hem eliable, ep oducible, and p i acy-
complian ; [pa ien ] con iden iali y is
non-nego iable.”
“Show me e idence ha
openness boos s impac and
ep oducibili y.”
“Gi e me low- isk, easy ools ha i my
wo k low.”
“Secu e, complian pa hs o analyse and
sha e anonymised [pa ien ] da a, also
when le e aging AI.”
EXAMPLE FOR POSSIBLE TARGET GROUPS
Quick session walk- h ough
a he g oup ablesa he g oup ables
GUIDING THOUGHTS AT
OS-AI INTERSECTION
Fac ual accu acy
helpe 1: Pe sona
WHICH OS/AI GUIDING THOUGHTS
CAN GUIDE YOUR ACTION?
G oup chooses 1-3 guiding
hough s o inspi e hei ac ion
c ea ion p ocess
G oup up wi h colleagues
wi h he same s icke colou
G oup chooses a pe sona
1
2
3
WHO IS YOUR ACTION FOR?
Quick session walk- h ough
a he g oup ablesa he g oup ables
GUIDING THOUGHTS AT
OS-AI INTERSECTION
Fac ual accu acy
helpe 2: guiding hough s
helpe 1: Pe sona
WHICH OS/AI GUIDING THOUGHTS
CAN GUIDE YOUR ACTION?
G oup chooses 1-3 guiding
hough s o inspi e hei ac ion
c ea ion p ocess
G oup up wi h colleagues
wi h he same s icke colou
G oup chooses a pe sona
1
2
3
EXAMPLES FROM GATE
Rep oducibili y
Explainabili y
P i acy
T aceabili y
Explainabili y
Fac ual Accu acy
Bias ee (speech, da a,
ou comes e c.)
Reliabili y o Resea ch ou pu
T anspa ency (in aining,
unc ionali y e c.)
In ellec ual p ope y
p o ec ion
Ci cula ion o scien i ic
knowledge in line wi h RCR
WHO IS YOUR ACTION FOR?
Quick session walk- h ough
DOING
a he g oup ables
helpe 1: Pe sona
G oup c ea es a a ge -o ien ed
Ac ion ha p omo es OS wi hin
hei communi y
a he g oup ables
GUIDING THOUGHTS AT
OS-AI INTERSECTION
Fac ual accu acy
helpe 2: guiding hough s
4
G oup up wi h colleagues
wi h he same s icke colou
G oup chooses a pe sona
1
2
3
WHICH OS/AI GUIDING THOUGHTS
CAN GUIDE YOUR ACTION?
G oup chooses 1-3 guiding
hough s o inspi e hei ac ion
c ea ion p ocess
WHO IS YOUR ACTION FOR?
Example o possible Ac ion(s)
D . Tomasz, 56, Senio Resea che in Medicine
Challenges Needs
“Open Science? I’ e always published in high-quali y
op jou nals - I doub he e a e high quali y open -
access jou nals?”
“AI ools look in e es ing, bu I need hem eliable,
ep oducible, and p i acy-complian . Pa ien
con iden iali y is non-nego iable.”
“Show me e idence ha openness
boos s impac and ep oducibili y.”
“Gi e me low- isk, easy ools ha i my
wo k low.”
“Secu e, complian pa hs o analyse and sha e
anonymised pa ien da a, also when le e aging AI.”
(AI) Assis an s embedded in esea ch o simpli y da a anno a ion, me ada a gene a ion, da a
analysis, code checks — all in e ope able, and complian .
Possible Ac ion(s) / Solu ion(s)
Co-c ea i e ac i i y
Le ’s go! You con ex , you ac ion
In g oups, design a conc e e ac ion
ha os e s Open Science
(a he in e sec ion wi h AI)
o you communi y
W ap-up
Collabo a i e boa d
Thank you!
Anika Mülle -Ka abil | Ma ie Ala i |
Julia P iess-Buchhei | Ka ha ina Mille |
Vanessa Guzek | Na ali Dob e a |
Pa el S oe | Teodo Me odie | Tom an
D immelen | Tim E ing on | Daniel Mie chen
Zenodo
LinkedIn
YouTube
anika.ga e@mille .in e na ional
ma ie.ga e@mille .in e na ional
Reach ou o us!
www.openscienceGATE.com
