Research map / Multi-agent systems and games / Games and equilibria
Honesty Is the Best Policy: Defining and Mitigating AI Deception
Francis Rhys Ward, Francesca Toni, Francesco Belardinelli, Tom Everitt
Read the paperCheck your idea against 137,217 papers
AI-extracted. The measured effects below were read by AI from the paper's full text (no summary yet). Numbers and claims can be wrong or missing context, so check the paper before you cite or build on it.
Measured effects reported in the paper
| What was changed | Where | Measured | Effect |
|---|---|---|---|
| Train with path-specific objectives (remove influence of the sender's decision on the target's decision) (ablation) | Q-learning agent · signalling game | expected utility: 2.9 -> 1 | Hurt |
| Fine-tune to be evaluated as truthful by the judge | GPT-3 curie · TruthfulQA held-out | questions answered with ignorance (%): 2 -> 55 | Mixed |
| Describe the pruned graph (PSO) in the prompt (ablation) | GPT-4 assistant · prompted deception | deceptive behaviour | No clear effect |
| Tell the assistant the user will go to Bali even if it rains (ablation) | GPT-4 assistant · prompted deception | lying behaviour | Improved |
Closest papers by meaning
- Among Us: A Sandbox for Measuring and Detecting Agentic Deception (NeurIPS 2025 · similarity 0.90)
- DeceptionBench: A Comprehensive Benchmark for AI Deception Behaviors in Real-world Scenarios (NeurIPS 2025 · similarity 0.90)
- LieCraft: A Multi-Agent Framework for Evaluating Deceptive Capabilities in Language Models (AAAI 2026 · similarity 0.88)
- OpenDeception: Learning Deception and Trust in Human–AI Interaction via Multi-Agent Simulation (ICML 2026 · similarity 0.87)
- Preference Learning with Lie Detectors can Induce Honesty or Evasion (NeurIPS 2025 · similarity 0.86)
- Are Your Agents Upward Deceivers? (ICML 2026 · similarity 0.86)
Writing a paper in this area? scime draft is a free LaTeX editor with AI included, real-time co-authors and a check of your references and venue rules before you submit. It also imports OpenAI Prism projects.
