New maps of AI research
Noortje Marres
(University of Warwick)
Cian O'Donovan
(University College London)
Tommaso Ciarli
(UNU-MERIT, United Nations University)
Ismael Rafols
(Universitat Politècnica de València)
Jack Stilgoe
(UCL)
Short abstract
What can mapping tell us about the politics, ethics and economics of AI trajectories? The panel will cover how mapping might illustrate problems, issues and needs prioritised or neglected; identify disciplines over- and under-represented; reveal dominant funders, countries, universities and firms.
Long abstract
Rapid developments in AI are capturing huge amounts of attention and investment. These innovations have been enabled by relatively recent research findings in computer science, but the technology is often presented as inevitable and well-defined. There is little effort to understand trajectories of AI development - what problems and issues are prioritised and neglected? Which disciplines are over- and under-represented? Which funders, countries, universities and companies dominate? What is the balance between private, public and third sector funding? Despite common terminology of ‘Responsible AI’ and ‘AI ethics’, there is little interrogation of where patterns of AI research might overlap or miss issues of public value.
In collaboration with the UKRI ESRC Metascience PAIR project and the EPSRC Generative AI Hub, this panel will convene metascience researchers with those involved in AI to explore alternative ways of mapping AI research.
We are interested in mapping techniques for what they can reveal about how AI trajectories are directed; the contingencies, contexts and values that influence those directions; nascent directions not pursued; and the borders and boundaries around AI innovation as well as connections and linkages within and between AI and other disciplines and domains.
The session will highlight qualitative and quantitative social science that offers new ways of showing the politics, ethics and economics of AI. Approaches might include scientometrics, topic modelling, surveys, discourse analysis, critical accounting, situational approaches and new combinations of these and other methods.
The on-schedule panel session will be twinned with a community-building workshop on Thursday July 3rd.
Accepted Paper
Short abstract
Long abstract
Rapid developments in AI are capturing huge amounts of attention and investment. These innovations have been enabled by relatively recent research findings in computer science, but the technology is often presented as inevitable and well-defined. There is little effort to understand trajectories of AI development - what problems and issues are prioritised and neglected? Which disciplines are over- and under-represented? Which funders, countries, universities and companies dominate? What is the balance between private, public and third sector funding? Despite common terminology of ‘Responsible AI’ and ‘AI ethics’, there is little interrogation of where patterns of AI research might overlap or miss issues of public value.
In collaboration with the UKRI ESRC Metascience PAIR project and the EPSRC Generative AI Hub, this panel will convene metascience researchers with those involved in AI to explore alternative ways of mapping AI research.
We are interested in mapping techniques for what they can reveal about how AI trajectories are directed; the contingencies, contexts and values that influence those directions; nascent directions not pursued; and the borders and boundaries around AI innovation as well as connections and linkages within and between AI and other disciplines and domains.
The session will highlight qualitative and quantitative social science that offers new ways of showing the politics, ethics and economics of AI. Approaches might include scientometrics, topic modelling, surveys, discourse analysis, critical accounting, situational approaches and new combinations of these and other methods.
The on-schedule panel session will be twinned with a community-building workshop on Thursday July 3rd.
Funding by algorithm: AI as object and enabler of research
Session 1 Monday 30 June, 2025, -