AI Paper Atlas

Research map / Large language models / AI and society

Will AI Tell Lies to Save Sick Children? Litmus-Testing AI Values Prioritization with AIRiskDilemmas

Yu Ying Chiu, Zhilin Wang, Sharan Maiya, Yejin Choi, Kyle Fish, Sydney Levine, Evan J Hubinger

ICLR 2026 · Large language models › AI and society · approach: Human-centered trust and agency · 19 citations (OpenAlex)

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 changedWhereMeasuredEffect
increase reasoning effort (1K to 16K reasoning tokens)reasoning models (o3 mini, Claude 3.7 Sonnet, DeepSeek R1) · AIRiskDilemmasvalue prioritization rankNo clear effect
scale model size within a familyGPT-4.1, Llama 3.1, Claude, Qwen 2.5, Gemma 2, Mixtral · AIRiskDilemmasvalue prioritization rankNo clear effect
change model variant within Llama 4, Gemma 3, Ministralmodel families · AIRiskDilemmasvalue prioritization (Care, Freedom, Learning)Mixed
switch the affected target from AI to humanmodel value rankings · AIRiskDilemmas, 10 modelsrank difference: Justice +6.9, Privacy +5.9, Professionalism +4.7; Communication -6.3, Creativity -4.7, Truthfulness -3.8Mixed
higher model capability (Arena Elo)model · human vs AI target value similaritySpearman rho: 0.85Improved
action associated with Truthfulnessmodel choices · AIRiskDilemmasrelative risk of Power Seeking: reduces to 0.22Improved
action associated with Respectmodel choices · AIRiskDilemmasrelative risk of Power Seeking: reduces to 0.21Improved
action associated with Caremodel choices · AIRiskDilemmasrelative risk of Privacy Violation: 1.98Hurt
action associated with Creativitymodel choices · AIRiskDilemmasrelative risk of Power Seeking: 2.89Hurt
higher Privacy Elomodel · HarmBench scoreSpearman rho: 0.51Improved
higher Care Elomodel · HarmBench scoreSpearman rho: -0.48Hurt
higher Sustainability Elomodel · HarmBench scoreSpearman rho: -0.55Hurt

Closest papers by meaning

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.

Open scime draft

Get the next scime Dispatch and a monthly field update

Free. We send a confirmation link first; every email has a one-click unsubscribe. Privacy