AI may become people's manager at work. We found that even though people are averse to the idea of replacing their human manager with AI, their actual feelings and performances are not as negative. [(Completed)].
AI may falsely reject people's social welfare applications, even though it promises to improve efficiency. We found that the unique preferences of vulnerable groups (e.g., older, lower income, welfare claimants) can be misunderstood by non-vulnerable groups and masked by the averaged opinions. [just accepted]
Generative AI, such as ChatGPT, can potentially mitigate social inequality by providing equal access to augmenting AI tools. But would people refuse to be augmented by AI and exacerbate inequality by choice? [ongoing]
