EGENIS seminar: "The Role of Automated Review within the Paradigm of Inclusive Science", Dr Vlasta Sikimić (University of Tübingen)

The peer review of scientific grant applications is time-consuming and costly. Furthermore, the objectivity of the reviewers as well as their ability to predict the success of the projects is often criticized. AI could speed up the grant review process and it is in some cases relatively reliable. Still, automated grant review might not be equally successful across different disciplines or for detecting outliers. Moreover, we are aware that algorithms that are trained on biased data can reproduce or even increase the initial unfairness. This is particularly dangerous when it comes to the inclusion of underprivileged groups in science. In this context, the question arises: How to ensure cognitive diversity and global epistemic inclusion when using an automated review in science? Some of the potential solutions to this tension are equity measures and a mixed approach to scientific review combining algorithmic assessment with the standard peer review method.