Author: admin

Under the code name “hera2”, a Prolog implementation of HERA is available from now. It can be obtain from github here: https://github.com/existenzquantor/hera2

The HERA machine ethics Python library is now maintained on GitHub: https://github.com/existenzquantor/ethics . This move saves us from some double bookkeeping, and provides HERA users with a more convenient way of using and learning about HERA.

Under “software”, we have added a new tutorial on using the new explanation API of HERA. It can be directly accessed via http://www.hera-project.com/explainable-causal-agency-models/

We have recorded a new demo video showing Immanuel having a discussion with Laura about the moral permissibility of a white lie.

Our paper on “Moral Permissibility of Action Plans” (Authors: Felix Lindner, Robert Mattmüller, Bernhard Nebel) has been accepted for publication in the Proceedings of AAAI-19.

Laura Wächter and Felix Lindner got their paper “An Explorative Comparison of Blame Attributions to Companion Robots Across Various Moral Dilemmas” accepted for the Human-Agent Interaction 2018 conference taking place in Southampton this year.

Hanna Stellmach has presented our Paper “Perception of an Uncertain Ethical Reasoning Robot: A Pilot Study” at the GI conference “Mensch und Computer 2018” in Dresden.

A short tutorial on using the new format called Moral Planning Domain Definition has been put online. It can be checked out here. The format has first been used in our Paper Moral Permissibility of Action Plans (PDF). It generalizes our former approach in that it allows sequences of actions, i.e., action plans, to be specified. It supports events (a.k.a. timed exogeneous actions) and conditional effects to describe more complex iinteractions between the agent’s actions and nature’s actions. Moreover, it supports descriptions of intrinsic moral value of actions, utility of facts, and positive and negative effects on particular moral patients. This way, descriptions in the MPDD format can be evaluated using all ethical principles currently implemented in HERA.

Apart from that, a tutorial on Kantian causal agency models has been added as an introduction to a Kantian variation of our original utility-based causal agency models. However, in the long run, the MPDD format as a generalization of both former kinds of causal agency models will be the preferred format to be used and considered in further developments of HERA.

• A generalization of the HERA framework to action plans (rather than single actions) has been presented at the ICAPS workshop XAIP: http://icaps18.icaps-conference.org/xaip/
• The formalization of Kant’s Humanity formula has been presented at DEON 2018: https://deon2018.sites.uu.nl/

Our paper “The HERA Approach to Morally Competent Robots” has been accepted for publication in the Proceedings of the 2017 IEEE/RSJ International Conference on Intelligent Robots and Systems (IROS 2017). See you in Vancouver!:-)