| Michael Rovatsos (mrovatso) |
| Fri 27 Mar 2015, 12:00 - 12:25 |
| Informatics Forum (IF-2.33) |
If you have a question about this talk, please contact: Dimitrios Diochnos (ddiochno)
In this talk, I will present a tentative programme of research to investigate novel methods for avoiding and resolving conflicts that may arise between artificial agents and human users. The proposed research takes the view that agents themselves have to become risk- and conflict-aware if we are to avoid unforeseen harm and damage to humans, and suggests a number of concrete research topics that are worth considering in this context. This talk will be followed by a shorter non-technical presentation on "safe AI" as part of the Jamboree (in G.07, from 12:30).