Può l'IA negoziare un aumento di stipendio per un dipendente umano in una riunione aziendale simulata ?

I sistemi di intelligenza artificiale vengono sempre più utilizzati per assistere o addirittura guidare compiti delle risorse umane. Queste negoziazioni richiedono la lettura dei segnali sociali, la comprensione delle dinamiche di potere e degli interessi a lungo termine delle relazioni. I moderni LLM gestiscono scenari di role-playing con notevole adattabilità. Navigano obiettivi mutevoli e punti di pressione emotiva senza istruzioni dirette.

Background

AI systems are increasingly used to assist or even lead human resources tasks. These negotiations require reading social cues, understanding power dynamics and long-term relationship stakes. Modern LLMs handle role-playing scenarios with impressive adaptability. They navigate shifting goals and emotional pressure points without direct instructions.

Current AI systems can role-play a simulated salary negotiation, parsing candidate talking points, market benchmarks, and company constraints to generate context-appropriate responses in real time. They can adjust tone, employ persuasive strategies, and even anticipate counterarguments, though they do not possess true intent or legal authority to finalize compensation changes. Benchmarks from controlled studies show these systems achieve parity with novice HR assistants on scripted negotiation scenarios, but still lag behind experienced managers when handling unscripted emotional or organizational complexities. Integrations typically rely on pre-loaded corporate policies and require human-in-the-loop approval before any binding decisions. — Enriched May 12, 2026 · Source: MIT Sloan Management Review