During the explanatory process, the explainer generally tries to find out whether the person he or she is talking to is following the explanation. To do this, the explainer must perceive non-verbal signals from the other person and adjust their explanation accordingly. Individual and situational factors can influence these social signals. Feedback signals can differ, for example, for people with psychological conditions compared to those without, as well as for people in stressful situations as opposed to those in neutral situations. This diversity of signaling should also be taken into account by artificial intelligence (AI) systems when explaining topics to humans, as project investigator Junior Professor Dr. Hanna Drimalla herself explains.

These issues and more are being explored in the TRR sub-project A06 by researchers from the fields of psychology and computer science. A06’s key research objectives include:

1. Developing an understanding of how social signals from the explainee are influenced by intra-individual (e.g. psychological conditions) and inter-individual (e.g. stress) factors in an explanation.
2. Investigating the extent to which automated monitoring models developed using data from people without psychological conditions in neutral situations are generalizable to inter- and intra-individual varieties.
3. Building an explanatory AI system that can deal with these effects. Following a co-constructive approach, it will work together with the explainee to develop an understanding of social signals.

Research areas: Psychology, Computer science