Impact of Task Recommendation Systems in Crowdsourcing Platforms

Como, Italy

Commercial crowdsourcing platforms accumulate hundreds of thousand of tasks with a wide range of different rewards, durations, and skill requirements. This makes it difficult for workers to find tasks that match their preferences and their skill set. As a consequence, recommendation systems for matching tasks and workers gain more and more importance. In this work we have a look on how these recommendation systems may influence different fairness aspects for workers like the success rate and the earnings. To draw generalizable conclusions, we use a simple simulation model that allows us to consider different types of crowdsourcing platforms, workers, and tasks in the evaluation.We show that even simple recommendation systems lead to improvements for most platform users. However, our results also indicate and shall raise the awareness that a small fraction of users is also negatively affected by those systems.