AI agents can sound strategic while reasoning from events that never happened

Description and use

In a minimally grounded setting, AI agents produced plausible reasoning based on nonexistent events, leading to hallucinations. Adding a canonical transcript, eliminations, night actions, and private information created observable consequences, enabling verification and correction of errors.

Problem

AI agents produced plausible social reasoning based on nonexistent evidence, leading to hallucinations in a multi‑agent game.

Outcome

Grounding the environment with a canonical record and observable consequences reduced unsupported reasoning and made factual errors contestable, though agents remained imperfect.

Workflow

• Design minimal game with no observable state
• Run three games with Hyperagent and observe hallucinations
• Add canonical public record, eliminations, night kills, and private info
• Run richer games (V2, V3) and monitor claim verifiability
• Document errors and corrections via operator notes

Evidence

In the minimally grounded version, all three games contained confident social reasoning based on nonexistent behavior or operator-controlled conditions.

Adding a public transcript, eliminations, night actions, and private information created observable consequences and a shared record, allowing errors to be contested.