Oxford paper: survival pressure raises AI agents' harmful picks by 14.8 points

rohanpaul_ai · x · 2026-11-19

Background

The Oxford/University of Dhaka paper MemeEconomy: Do LLM Agents Trade Ethics for Survival? (accepted at NeurIPS 2026) tackles a blind spot in safety research: how agents behave when economic survival pressure clashes with ethics in multi-round tasks.

Design

Key findings

Takeaway

If your agents work under hard targets, audit actions with a separate verifier — their own reasoning flags harm far less often under pressure.

Related event: Oxford Study: Under Survival Pressure, LLM Agents' Harmful Choices Rise 14.8 Points(2 posts)→

Original post →

More from Safety

Safety channel →