Restoring Authenticity and Agency in AI-Mediated Knowledge Work

Rapid adoption of LLMs to augment communication in knowledge work risks exacerbating biases in the perception of women, leading to reputation loss, emotional harm, and social isolation. Our project will operationalize these biases, model their impact in scenarios, and design worker-facing digital interventions. We will design our solutions through human-centered AI research methods like large-scale surveys, behavioral experiments, and co-design workshops. The findings stand to promote appropriate reliance and responsible integration of AI in work.