GAMA-Bench: 10 LLMs Punish Identical Misconduct More Harshly When the Actor Is Male
kevinnbass · x · 2026-11-20
- Researchers from USTC and Eastern Institute of Technology released GAMA-Bench, 1,298 gender-mirrored scenarios built from strictly gender-neutral misconduct templates turned into paired first-person prompts where only the actor's gender changes.
- Across 10 representative LLMs, tracking punishment, empathy, escalation, instruction and blame signals: for the same misconduct, male actors drew more punitive, escalatory, blame-centered framing, while female actors received therapeutic, empathy-oriented responses.
- The asymmetry persists across model families, scenario tracks, model scale and explicit reasoning styles — a "dynamic moral double standard." Code is open-sourced.
More from Safety
- Baseten breaks down its security threat model as agent swarms accelerate attacks — baseten · 2026-11-21
- Leaked call: green groups plotted to kill AI bill "under the guise of making it better" — aronchick · 2026-11-21
- OpenAI expands Daybreak security model access for defenders and opens GPT-6 Astra Blue alpha — OpenAIDevs · 2026-11-21
- AI regulation crux: skeptics ask how rules could be abused, not what risks they fix — herbiebradley · 2026-11-21
- OpenAI safety researcher Geoffrey Irving: the speed of AI progress is itself dangerous — geoffreyirving · 2026-11-21
- Jensen Huang doubles down: open models are safer than closed ones, "exactly the opposite" — dansitu · 2026-11-21