AbstractView references

We are witnessing a dramatic migration of society from the real world into virtual worlds (online systems) and face-to-virtual worlds (hybrid systems and events). Apart from the conventional e-gov and e-dem systems, virtual worlds (social media or other forms of online systems) become increasingly used as tools for promoting (or suppressing) democratic values and calls for changes. There is significant research on designing more efficient systems for practising e-democracy and (digital) ethics [1] [2]. However, fewer examples exist when it comes to mechanisms capable of managing democratic and ethical values in neutral digital systems whose main purpose is not related to e-gov and e-dem [3]. Together with a large stratum of society, power structures also migrate to virtual worlds. Therefore, there is a strong need to successfully migrate and enable democratic and ethical principles - one of the oldest and the most important values of a society.We propose a paradigm shift in system design to make online systems more transparent and open to external evaluation by neutral domain experts and institutions through visual workflows and behavioral processes. One implementation of such an approach is the ColaboFlow workflow framework, together with an associated domain-specific framework - Democracy Framework. We argue for the benefits of such an approach especially in the face-to-virtual context, which has become the primary modus operandi for many community-related events during the COVID- 19 era. We demonstrate ColaboFlow and Democracy in the real context of Wikipedia by analyzing its community and users, i.e. knowledge workers, and their behavior. We analyze and evaluate the Wikipedia community's engagement (motivation) and participation in collaborative work in the relationship in terms of democratic and ethical community behaviors. We observe the types and quantities of the Wikipedia policies that knowledge workers refer to in the process of deliberative democracy, and correlate these findings with user behavior and motivation for contributing within the online ecosystem. We propose a two-tier model where we detect system anomalies at the (i) high, statistical and resource-cheap level, and understand and semi-automatically normalize (heal) the recognized anomalies at the (ii) low, behavioral and resource-demanding level. © 2021 IEEE.