top of page
9894067_1182_edited.jpg

Research Matters

Examining non-human agents, such as non-embodied artificial intelligence systems, chatbots and autonomous robots, requires many choices when it comes to theories and methods.

robots-are-also-used-logistics-transportation-generative-ai.jpg

Anthropomorphism

The research project Interactions with Artificial Humans comprises several studies in which assumptions of anthropomorphism play a significant role.

The Wizard-of-Oz approach

The current state of non-human agents, such as robots, is that many are unable to do all the things that a researcher wants them to do in a study. One way to deal with this is the Wizard-of-Oz approach.

26986885_v882-aew-51-e.jpg
scientist-analyzing-medical-sample-with-modern-microscope-generative-ai.jpg

Experiments

Many of the studies in Interactions with Artificial Humans employ an experimental method. That is to say, a stimulus non-human is manipulated by the researcher so that it exists in different versions, and participants are randomly allocated to being exposed to one of these versions. 

Measuring participants' reactions

We humans can react to non-human agents in copious ways, and this is mirrored in research on robots, chatbots and virtual agents: a wide range of reactions are indeed measured.

national-mathematics-day-generative-ai.jpg
Magnus Söderlund takes a robot outdoors.jpg

Stimuli

In order to examine how we humans react when we are interacting with non-human agents, you must somehow expose human participants for non-human agents. This can be done in several different ways – and it is indeed done in different ways in existing research.

bottom of page