- Understanding the motivations underlying the human inclination to automate tasks is vital to developing truly helpful robots integrated into daily life. Accordingly, we ask: are individuals more inclined to automate chores based on the time they consume or the feelings experienced while performing them? This study explores these preferences and whether they vary across different social groups (i.e., gender category and income level). Leveraging data from the BEHAVIOR-1K dataset, the American Time-Use Survey, and the American Time-Use Survey Well-Being Module, we investigate the relationship between the desire for automation, time spent on daily activities, and their associated feelings - Happiness, Meaningfulness, Sadness, Painfulness, Stressfulness, or Tiredness. Our key findings show that, despite common assumptions, time spent does not strongly relate to the desire for automation for the general population. For the feelings analyzed, only happiness and pain are key indicators. Significant differences by gender and economic level also emerged: Women prefer to automate stressful activities, whereas men prefer to automate those that make them unhappy; mid-income individuals prioritize automating less enjoyable and meaningful activities, while low and high-income show no significant correlations. We hope our research helps motivate technologies to develop robots that match the priorities of potential users, moving domestic robotics toward more socially relevant solutions. We open-source all the data, including an online tool that enables the community to replicate our analysis and explore additional trends at https://robin-lab.cs.utexas.edu/why-automate-this/ -
+ What tasks do we want robots to handle? Are those preferences based on saved time or feelings we associate with the tasks? This study uncovers how feelings shape robot automation preferences, revealing insights across gender and income groups. Using data from BEHAVIOR-1K and the American Time-Use Survey, we show that feelings(Happiness, Stress, Meaningfulness), not time spent, drive automation desires. We open-sourced the data + built an interactive tool