A novel approach, AID, that combines Large Language Models (LLMs) and differential testing to efficiently generate test cases that can identify defects in plausible programs.
