Human Edge of AI: Assistant Professor Solène Delecourt

Why It Matters

Understanding AI bias in compensation advice is crucial for businesses aiming to close gender pay gaps and for professionals to rely on authentic negotiation skills.

Key Takeaways

• Generative AI often embeds gender bias in negotiation advice.
• AI may suggest lower salaries for older women than men.
• Test prompts with multiple personas to spot biased recommendations.
• Real‑world negotiation skills still outweigh AI‑generated suggestions significantly.
• Students should prioritize human interaction, tone, and relationship building.

Summary

Assistant Professor Solène Delecourt of Berkeley Haas warns that generative AI tools, increasingly used by MBA students for negotiation prep, carry hidden biases.

Her research shows AI often infers lower experience and competence for women—particularly older women—leading to advice that recommends smaller salary figures compared with male counterparts.

She illustrates this with examples of AI suggesting reduced compensation and urges students to run identical prompts across varied personas to detect divergent recommendations.

Delecourt stresses that despite AI convenience, effective negotiation still depends on in‑person cues, tone, timing, and relationship building, making human skill development essential for closing the gender pay gap.

Original Description

Here's what to know before you use AI to prepare for a negotiation: Insights from Assistant Professor Solène Delecourt, who teaches negotiation MBA students and whose research has documented the deep biases embedded in AI.

#HumanEdgeofAI #HaasAI #BerkeleyHaas #UCBerkeley #Innovation #Leadership