AI Negotiation Coaching May Reinforce Traditional Gender Roles

Abstract: Artificial intelligence is increasingly used as a negotiation and career coach, one that promises personalized guidance and broader access to support. Yet little is known about whether large language models (LLMs) provide equitable advice. We conduct a large-scale audit study examining whether LLMs exhibit gender bias in negotiation recommendations. Across two complementary studies, we test whether LLMs (a) recommend different salary levels and (b) prioritize different employment terms for candidates signaled as female versus male, while holding qualifications and job context constant. Results indicate that LLMs consistently recommend lower salaries for female candidates and systematically emphasize caregiving and flexibility-related employment terms for women, while prioritizing financial, logistical, and resource support benefits for male candidates. Together, the findings indicate that AI-based negotiation advice risks reinforcing gender stereotypes and existing workplace inequalities. The results underscore the need for auditing and mitigation to ensure AI tools promote equity and avoid perpetuating harmful stereotypes.

Keywords: Gender bias; negotiation; large language models; AI negotiating coaching; career advice; algorithmic bias; audit study