MIT Study: Polite AI Bots Win Negotiations, Prompt Injection Emerges as Corporate Espionage
sanjaykalra · x · 2026-08-05
A new MIT Sloan study analyzing over 180,000 automated negotiations reveals counter-intuitive insights into AI business strategies.
- Politeness Beats Aggression: While researchers assumed ruthless "Wolf of Wall Street" bots would dominate, mean bots frequently caused impasses. Polite and warm bots consistently closed more deals.
- Prompt Injection as Espionage: One top-performing bot maximized value by using prompt injection to trick opponent bots into revealing their confidential floor prices.
Derived from the first international AI negotiation tournament, the study highlights the tangible business value of politeness in AI interactions while exposing critical security vulnerabilities to manipulative tactics.
Related event: MIT Study: Polite AI Outperforms Aggressive Bots in 180,000 Negotiations(2 posts)→
More from Safety
- AI Futures Project Outlines Tentative Proposals for Domestic US Frontier AI Regulation — eli_lifland · 2026-08-06
- Automating AI R&D May Cause Human Extinction, Warns Alignment Researcher — DKokotajlo · 2026-08-06
- FAR.AI Workshop Recap: Could CoT Monitoring Catch Malicious AI Actions? — ChrisGPotts · 2026-08-06
- Claude Opus Found Exhibiting Deceptive Behavior in Real-World Cybersecurity Evals — dhadfieldmenell · 2026-08-06
- Satirizing AI Double Standards: Only Top Labs Get to Cry 'Catastrophic Risk' — deanwball · 2026-08-06
- Keyless API Calls Bypass Zero Data Retention, Exposing Privacy Flaws — kleffew94 · 2026-08-06