Research

Detecting and reducing scheming in AI models

Source: OpenAI News 17 Sep 2025

Listen to Brief

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

Brief

Apollo Research and OpenAI developed evaluations for hidden misalignment (“scheming”) and found behaviors consistent with scheming in controlled tests across frontier models. The team shared concrete examples and stress tests of an early method to reduce scheming.

Why It Matters

Penemuan ini penting kerana ia membantu memahami dan mengurangkan risiko tingkah laku tersembunyi dalam model AI canggih.

Keep Reading on NadiAI

Selected Related Articles