Accountability vs. Safety in Permissioned Decision Systems
Accountability vs. Safety in Permissioned Voting and/or Decision Systems
Convener: Dave Sandford
Notes-taker(s): Dave Sandford
Discussion notes, key understandings, outstanding questions, observations, and, if appropriate to this discussion: action items, next steps:
We discussed identity and how a legal name or even long standing pseudonym can provide a reputation history and accountability for deliberations within a decision support system. Exposing legal names and valued pseudonyms can also allow attacks and reduce safety. Assumes a system which ensures participants are vetted and cannot use more than one pseudonym.
A concentric circle permissioned decision model was discussed where all participants are vetted - and can interact as:
- anonymous vetted (anonymous to other participants but vetted as a valid participant, could be implemented via ring signatures, or other means)
- legal name
A concentric circle model was discussed where:
Outer circle - Purpose: discussion, Constraint: all participants can be anonymous vetted
Next inner circle - Purpose: voting or other human consensus process, Constraint: participant would have to identify themselves with their 'vote' at least pseudonmously
Innermost circle - Purpose: roles related to maintaining the decision making system process, Constraint: participant would have to identify by legal name
There might be an amount of time to be able to build the reputation to move from outer to inner circles. The intent would be to minimize barrier of entry to outer circle, and at each level balancing the cost of entry with the cost of safety.