Research

AI-written critiques help humans notice flaws

Source: OpenAI News 13 Jun 2022

Listen to Brief

AI audio in English, based on the NadiAI brief and original source.

Brief

We trained “critique-writing” models to describe flaws in summaries. Human evaluators find flaws in summaries much more often when shown our model’s critiques. Larger models are better at self-critiquing, with scale improving critique-writing more than summary-writing. This shows promise for using AI systems to assist human supervision of AI systems on difficult tasks.

Why It Matters

Pendekatan ini boleh meningkatkan ketepatan penilaian manusia terhadap sistem AI dalam tugasan kompleks.

Keep Reading on NadiAI

Selected Related Articles