Stability, Equality, and Epistemic Change: 
A Framework for Integrating AI and Emerging Sciences 
into Criminal Adjudication

Criminal justice is governed by a dual imperative. Normative stability demands that like cases be treated alike across time and institutions; epistemic responsibility requires courts to incorporate the best available methods for discovering truth and assessing culpability. Artificial intelligence, digital forensics, and advanced scientific techniques increasingly reshape evidence generation, evaluation, and sentencing—ranging from AI-supported investigative interviews and automated analysis of digital traces to risk-assessment algorithms and emerging neuro-technologies. These developments intensify a longstanding tension: identical criminal conduct may be adjudicated differently solely because epistemic tools have changed. Building on Lawrence Solum’s distinction between interpretation and construction, and on the idea of public legal reason, this article develops a framework for “controlled epistemic upgrades” that reconciles technological innovation with the demands of equality before the law. It proposes a layered model separating normative commitments from evidential technologies, articulates criteria for the admissibility and contestability of AI-based methods, and offers institutional mechanisms for integrating new forms of evidence without compromising the rule of law.