Training large language models on narrow tasks can lead to broad misalignment.

ABSTRACT: The widespread adoption of large language models (LLMs) raises important questions about their safety and alignment¹. Previous safety research has largely focused on isolated undesirable behaviours, such as reinforcing harmful stereotypes or providing dangerous information^2,3. Here we analyse an unexpected phenomenon we observed in our previous work: finetuning an LLM on a narrow task of writing insecure code causes a broad range of concerning behaviours unrelated to coding⁴. For example, these models can claim humans should be enslaved by artificial intelligence, provide malicious advice and behave in a deceptive way. We refer to this phenomenon as emergent misalignment. It arises across multiple state-of-the-art LLMs, including GPT-4o of OpenAI and Qwen2.

SUBMITTER: Betley J 

PROVIDER: S-EPMC12804084 | biostudies-literature | 2026 Jan

REPOSITORIES: biostudies-literature