From Information to Incarnation: Reviewing Half a Century of AI/Data Activities to Realize New Materials From Data Using the Materials Innovation Crucible

ABSTRACT The use of informatics for materials design has long promised revolutionary advances through data‐driven discovery; but the untrustworthiness of the available data continues to undermine progress. Indeed, materials knowledge remains fragmented across disciplines and organizations; collaboration faces structural barriers, and the gap between digital representations and experimental reality continues to widen. In this paper, I argue that much of this confusion stems from a failure to distinguish between two fundamentally different levels of persistent identification: One level addresses matter—substances that are characterizable by well‐defined scientific concepts, such as crystal structure—using intentional definitions specifying membership criteria rather than enumerated instances to enable stable entity identification; the other level addresses materials, such as alloys for jet engine turbine blades, whose microstructures, properties, and behaviors evolve dynamically under service conditions, making the persistent identification of individual facts inherently context‐dependent. Conventional materials database approaches have conflated these levels, assigning identifiers without determining whether they reference conceptually defined sets or individual context‐bound facts. This conflation, rather than mere technical or economic limitations, lies at the heart of the current crisis in materials data reliability. Drawing on concrete examples from nuclear reactor and high‐temperature superconductor materials, I demonstrate how such critical context can be lost during abstraction, making the realization of materials “from information into incarnation” nearly impossible. I connect these limits to their theoretical foundations in Hilbert’s formalism, Gödel’s incompleteness, and Turing’s undecidability. To support the creation of a trustworthy common platform for materials design, I propose the materials innovation crucible (MIC) framework, which addresses the identified challenges using: (1) Traceable knowledge pathways that preserve context and uncertainty to explicitly distinguish set‐level data points from individual identifiers from individual facts, (2) incentive‐aligned sharing mechanisms, and (3) evolutionary discovery within transparently constrained spaces. Central to this approach is multi‐level knowledge representation that acknowledges the impossibility of complete formalization while maintaining rigor. Finally, I present MIC implementation strategies that can be evaluated by constructing databases for materials design, and outline pathways for broader adoption across materials science and engineering.