Human computation has come of age in recent times, as an increasing number of projects recruit thousands of volunteers to beaver away on relatively small chunks of a much larger project.

With human computation, there often appears to be a distinction between tasks humans can do and tasks that computers can do, and that boundary remains rather fixed.

Blurring the boundaries

A recent study suggests that this boundary should be blurred however, and that AI and the crowd can work rather effectively together.

The paper proposes a new future for human computation that combines both machine and human intelligence.  It paints a vision of a hybrid computational system that is capable of tackling the biggest and most intractable problems currently facing humanity.

They’re thinking of the kind of ‘wicked problems’ that have for so long defeated mankind and that tend to defy more traditional ways of doing things.

A hybrid approach

The authors believe that a well designed AI can significantly enhance our cognitive abilities and that when working together we are capable of forming highly capable collaborative networks.

For instance, AI may be able to effectively manage a crowd of human contributors so that the pool of participants is effectively coordinated.

For instance, such approaches are already being tested, albeit in small ways, in projects such as YardMap.  The researchers believe however that it can certainly be scaled up, especially in areas such as medical research.

This is an area where human computation has already achieved substantial results.  For instance, the We Cure ALZ project uses crowd based microtasking, and the team hope to shrink treatment discovery times from decades to a few years.

Suffice to say, with such a new area it has a long way to go, but there are strong signs that it could achieve significant results.  I’ve already written about the use of AI ‘managers’ in companies such as Uber and it will be interesting to see if it can cross over into scientific domains.