Quaquebeke et al. say regular employees will need to be multilevel managers of AI agents, setting their strategic goals (top level), designing workflows (mid-level), and quality control (hands-on operational). Costs might include ethical isolation and loss of meaning. Heavy AI users I talk to are working this way already. Those I talk to don’t report these downsides. But their early adoption reflects their technical over social interests. As more people begin to work this way I expect reports of these challenges will rise.
Source: https://www.sciencedirect.com/science/article/abs/pii/S1048984326000342
