Participatory Decision Ethics
There is an increasing awareness that participatory methods for decision-making (such as voting mechanisms) do not always reach the most desirable outcomes. Some of the most reported reasons are:
- The individual deviation from the aggregated result is too large leading to less commitment and disconnection from the majority decision (e.g. a large number of people in the U.S. currently carry badges or stickers indicating "Not my president", and in the UK indicating "I voted Remain").
- Socially accepted results are not necessarily morally acceptable.
Most tools for (online) decision making do not address this discrepancy sufficiently. In fact, these tools focus mostly on the argumentation process preceding the decision, giving little support to the acceptance process following the decision. In particular, we consider argumentation-based tools which provide participants in a debate the possibility to propose alternatives, support alternatives, add comments or justifications for alternatives, vote on alternatives. See www.airesis.eu, http://deliberatorium.mit.edu/login?english or https://mood.tbm.tudelft.nl/ for some examples
In this hackathon we are going to explore methods and tools to support decision-making. The following are some of possible directions for the work:
- Supporting different voting/aggregation possibilities (e.g. multiple rounds, voting on what one expects most will vote) and analysis of their effects on voting behavior. Visualizing the difference between own decision and the aggregate one (overall or grouped)
- Providing means to tag opinions and alternatives with values and to aggregate alternatives along value tags (e.g. option 1 promotes fairness, option 2 demotes privacy). Participants could choose from a list of values or use free text.
- (Semi-)automated extraction of personal values from word use (using e.g. Sentiment analysis tools or IBM Watson Tone Analysis app). See http://citeseerx.ist.psu.edu/viewdoc/download?doi=10.1.1.636.7168&rep=rep1&type=pdf for an example of this approach
- IBM Tone Analyser: https://www.ibm.com/watson/developercloud/tone-analyzer.html (emotion and sentence tone)
- IBM Personality Insights: https://personality-insights-livedemo.mybluemix.net/ (includes OCEAN and Schwartz values)
The aim is to arrive at a working prototype that implements (some of) these aspects. Participants are encouraged to use their prefered open source tools and reuse any suitable infrastructure. Resulting code will be made available open source for further exploration and use, e.g through Github.
For questions or further information, contact Virginia Dignum, firstname.lastname@example.org