Factors Affecting the Success of Social Machines

The final speaker in this Web Science 2016 session is Clare Hooper, whose interest is in 'social machines' as defined by Tim Berners-Lee: systems were people do the creative work, and machines take care of the administration. Social machines will exist in the context of problems to be solved; they may be created by single stakeholders (as in the case of Galaxy Zoo), while others arise in a more emergent fashion (from online communities).

But any social machine has more than one stakeholder, and these often have more or less strongly conflicting needs. The job market site Skills Planner involves the operating company, colleges, local councils, employees, and companies, for instance, whose interests will vary widely.

Important to certain social machines are geographical context, the reuse of goods, and offline meet-up opportunities; homophily of participants' social networks and similarity in timezones of participants may also be crucial. Questions about how the data are handled, and what geographic resolutions are supported to the machine may also be important for its success. Further, temporal contexts, the granularity of the data and metadata, and the balance between the accuracy of spatial data and currency of temporal data are similarly important. These are also matters of trust, security, and privacy.

Broader contextual changes are also an important factor in the success of social machines; changing technological frameworks and growing accessibility of new technologies may affect the continuing need for the social machine's services. Growth in the userbase may also necessitate further adjustments to how the social machine operates.