The Refutability Gap: Challenges in Validating Reasoning by Large Language Models
This work highlights critical methodological flaws in AI research on reasoning, aiming to improve scientific integrity for researchers and policymakers, though it is incremental as it builds on existing critiques of AI validation.
The paper argues that claims about large language models achieving new science or human-level intelligence lack scientific rigor because they fail to meet Popper's refutability principle, identifying methodological pitfalls like opaque training data and lack of reproducibility. It proposes guidelines for transparency and reproducibility to address these issues.
Recent reports claim that Large Language Models (LLMs) have achieved the ability to derive new science and exhibit human-level general intelligence. We argue that such claims are not rigorous scientific claims, as they do not satisfy Popper's refutability principle (often termed falsifiability), which requires that scientific statements be capable of being disproven. We identify several methodological pitfalls in current AI research on reasoning, including the inability to verify the novelty of findings due to opaque and non-searchable training data, the lack of reproducibility caused by continuous model updates, and the omission of human-interaction transcripts, which obscures the true source of scientific discovery. Additionally, the absence of counterfactuals and data on failed attempts creates a selection bias that may exaggerate LLM capabilities. To address these challenges, we propose guidelines for scientific transparency and reproducibility for research on reasoning by LLMs. Establishing such guidelines is crucial for both scientific integrity and the ongoing societal debates regarding fair data usage.