As collective intelligence gains prominence in the enterprise, interface designers must confront some new challenges. The controls on a word processor or spreadsheet are essentially deterministic. If you click "right align" you know the text is going to swing over to the right side of the page. The same user input always has the same output, barring bugs.

However, social (or rather "sociomimetic") technologies, which draw on intelligence from across a network to infer what information is relevant to each person, don't behave like this. They rely on increasingly complex algorithms which aren't directly comprehensible to the user and lack any direct link between cause and effect. A slight change in one parameter can lead to a radical and unpredictable change in output.

Conventional user-control principles need rethinking in these circumstances. There are too many factors and the interplay between them is too complex for direct control to work. One alternative is to take control away from users and let the system operate "by magic." But this leaves users feeling disempowered and unable to influence their environment.

The starting point for a solution lies in deep-rooted human nature. People require a mental model of how a tool works and a sense of control. To meet these needs in a non-deterministic system, it may be necessary to manufacture deliberate illusions about the system's function and controlability.

In this session, Charles Armstrong offers an ethnographer's perspective on how humans interact with complex systems and what this means for a new generation of sociomimetic technologies.