A Definition of Good Explanations and the Challenges Explaining LLM Outputs

Researchers are formalizing criteria for what constitutes a high-quality explanation for artificial intelligence outputs, addressing a longstanding philosophical debate recently revived by the rise of large language models. As AI integration grows in sensitive fields, establishing these standards is essential for determining how these systems communicate their reasoning to human users.