A core feature of digitization is the increasing delegation of decision-making functions to software systems. This applies in particular to cases in which human decision-makers have to make decisions under time pressure and uncertainty. Such decisions are typical for the security context (military, police, emergency medical response and rescue). It is to be expected that more and more digital systems will be used in this context to support the human decision maker or even make certain decisions.
This raises ethical questions about accountability and acceptance as technical systems are increasingly equipped with autonomous capabilities and operate in conditions where decision options become more diverse, but the decisions themselves raise critical ethical questions, e.g. the acceptance of collateral damage.
The aim of this project is to investigate how to ensure that digitized decision support and automated decision functions lead to an ethically acceptable result when a human decision maker has to make decisions quickly and under uncertainty. To this end, both normative analyses and experiments will be carried out. The project should lead to a system that will serve as a demonstrator for the systematic experimental investigation of this question and to generate data that can be used for the training of possible decision (support) systems in the future.
- Prof. Abraham Bernstein, DSI & Institut für Informatik der Universität Zürich
- PD Dr. Markus Christen, DSI & Institut für biomedizinische Ethik & Medizingeschichte der Universität Zürich
- Dr. Markus Kneer, Ethikzentrum der Universität Zürich
- MSc Suzanne Tolmeijer, Institut für Informatik der Universität Zürich
The project has been running since 2017 and is supported by a research contribution from armasuisse Research+Technology.