Article

Successful Automatic Model Discovery Can Produce False Mechanisms

2026-07-12

Abstract excerpt

<p>Automatic model discovery is appealing because it seems to offer what black-box prediction cannot: symbolic, executable accounts that can be read as candidate theories. That apparent advantage also creates a distinctive failure mode. When a predictive surrogate is expressed in familiar scientific language, it may be mistaken for the mechanism that generated the data. We tested this risk in a controlled risky-ch...

Topics

Open a Topic to create a Post that cites this publication.

Identifiers and source

Literature Corpus work: 79068b4f-57ae-539d-a093-84277f7907d9
DOI: 10.31234/osf.io/r46ux_v1

Open publication

Related research

Semantic proximity does not establish scientific evidence.

Click a neighbor to travelStep 1 · 12 closest

Successful Automatic Model Discovery Can Produce False MechanismsDOI 10.31234/osf.io/r46ux_v1Read

Computational Models Gone RogueSemantic proximity Read

On logical inference over brains, behaviour, and artificial neural networksSemantic proximity Read

Classifying 25 Misinterpretations of Statistical Tests: A Comparison of Six Large Language ModelsSemantic proximity Read

Recognition, Simulation, and Refusal: A Contamination-Aware Study of Classic Psychological Effects in LLM AgentsSemantic proximity Read

Artificial Cognition: How Experimental Psychology Can Help Generate Explainable Artificial IntelligenceSemantic proximity Read

In defense of epicycles: Embracing complexity in psychological explanationsSemantic proximity Read

Model shopping: how answer selection across AI models inflates scientific error ratesSemantic proximity Read

From Data to Discovery: How Foundation Models Are Accelerating Scientific Innovation with Generative AISemantic proximity Read

Mind the Gap: From Plausible to Valid Self-Explanations in Large Language ModelsSemantic proximity Read

Publishing Test-Ready Hypotheses When Experiments Are Out of ReachSemantic proximity Read

Statistical approximation is not general intelligenceSemantic proximity Read

Closed-Loop Scientific Discovery in the Behavioral SciencesSemantic proximity Read