PhD Candidate: Societal Implications of Artificial Intelligence at the Donders Centre for Cognition

Are you an aspiring researcher with an interest in the field of artificial intelligence? And would you like to delve into increasing the involvement of humans in decision-making processes? As a PhD candidate, you will be able to put your ideas to the test and push your boundaries. You do this in a collaborative, multidisciplinary and supportive work environment, with a diverse international staff. 

Human decision-making is increasingly shaped by Decision Support Systems. EU regulations request effective human oversight on decision-making. It is unclear whether humans can fulfil such a supervisory role for prolonged periods. The project will develop a prototype of a 'Reflection Machine' (RM) aimed at increasing human resistance against automation complacency. The prototype will be applied to the domain of chronic low back pain in collaboration with the Radboud university medical center and

Sint Maartenskliniek

.

The PhD project will develop an RM prototype that will provide feedback on joint human-DSS decisions, increasing the involvement of humans in the decision-making process. RMs improve human oversight by asking questions about the possible reasoning behind accepting or rejecting a recommendation. As a PhD candidate you will work on four subprojects: (1) Deriving design requirements from EU regulations such as the GDPR, the AI Act, and ethical codes such as developed by the EU High-Level Expert Group on trustworthy AI. (2) Specifying how RMs can intervene in decision-making by utilising insights from cognitive psychology and neuroscience regarding human decision-making while being supported by AI. This will involve working with the 'Nijmegen Decision Tool', a DSM developed for estimating treatment outcomes for chronic low back pain (CLBP) at the Radboud university medical center and Sint Maartenskliniek. (3) Developing a prototype implementation of an RM in the medical domain of CLBP on the basis of (1) and (2), such that its effects on automation complacency can be measured. The RM should produce output in the form of questions that prompt the physician to reflect on the decision more deeply. (4) Exploring the RM's functionality and user experiences, by investigating the effects of the RM prototype on the quality of decision-making (e.g. accuracy and efficiency), the potential interference with workflow, and satisfaction levels regarding the overall decision-making process. All projects will result in deliverables such as publications in international peer-reviewed journals and/or conference contributions. A small part of the available time (max. 0.2 FTE) will be devoted to teaching activities, such as guest lectures and Bachelor’s and Master’s thesis supervision, within the context of the AI educational programme.