When thinking about algorithms, cold lines of code and purely rational decisions may come to mind. However, this picture is incomplete. Numerous examples illustrate how human aspects shape algorithmic output (e.g., via biased training data). This study delves into how developers’ and users’ individual differences can influence algorithmic output, focusing on environmental and altruistic motivation. In an online survey, (N = 766) participants rated different emails on their likelihood of being spam as input for a hypothetical spam-filter algorithm. Participants’ environmental motivation was negatively correlated with classifying emails from environmental and humanitarian organizations as spam. Thus, individuals with a stronger environmental motivation rated the emails in such a way that the spam filter was biased toward the common good. However, altruistic motivation had no impact on the ratings. These findings suggest that environmental motivation extends beyond pro-environmental behaviors by also influencing prosocial behaviors, thus offering insights for developing sustainable algorithms.
https://doi.org/10.1145/3613904.3642404
The ACM CHI Conference on Human Factors in Computing Systems (https://chi2024.acm.org/)