Abstract: With the rapid development of information intelligence, the job tasks according to the profession of the labor force can be divided into routine tasks and non-routine tasks based on whether the tasks are easier to code and automate. It is a new perspective to understand the changes of income inequality from the reward differences of different tasks. Based on the data from the Chinese General Social Survey (CGSS) and the Occupational Information Network, this paper conducts a study of the reward differences of job tasks and the income inequality. The findings show that the denser the non-routine tasks in the occupations engaged by the labor force, the higher the returns they receive, and that the more intensive the routine tasks are, the lower the returns they receive. The phenomenon of income inequality within different types of work tasks is also not the same. In occupations with intensive non-routine tasks, the income inequality within the profession is higher; while in occupations with intensive routine tasks, the income inequality within the profession is relatively lower. In non-state-owned organizations and in the areas with advanced information technology, the higher returns of non-routine work tasks are more significant, and the problem of income inequality within the professions is severe. A counterfactual decomposition of the changes in the overall income inequality in China from 2002 to 2020 reveals that the difference in returns between non-routine and routine tasks is the main cause of overall income inequality, with an increasing effect on the overall income inequality. The structural changes in non-routine and routine tasks have little impact on the overall income inequality.

Key words: non-routine tasks, routine tasks, income inequality, information technology