AI predictive models do not make law deterministic

What the study found: AI-based predictive systems may enhance consistency in judicial decision-making, but the article argues they do not turn law into a deterministic system. The authors say predictive models produce probabilistic expectations shaped by past decisions and by temporal, institutional, and social dynamics.

Why the authors say this matters: The study suggests that artificial intelligence should be understood within the structural uncertainty of law, rather than as a way to overcome it. The authors conclude that predictive accuracy does not eliminate indeterminacy in law, even if it surpasses human expertise in specific tasks.

What the researchers tested: The article uses a conceptual and comparative analysis grounded in legal theory, philosophy of science, and socio-legal research. It examines whether contemporary AI-based predictive systems revive idealized models of legal determinacy using Dworkin’s Judge Hercules and Laplace’s Demon as reference points.

What worked and what didn't: The article states that predictive accuracy can be high enough to outperform human expertise in certain tasks. However, it also argues that this does not change law into a deterministic system, and that indeterminacy remains.

What to keep in mind: The abstract describes a theoretical analysis, not an empirical test of a specific AI system in court. It does not provide detailed limits, case examples, or measurements beyond the broad claim that predictive models are conditioned by temporal, institutional, and social dynamics.

Key points

• The article argues that AI predictive systems do not make law deterministic.
• It says these systems can generate probabilistic expectations from past decisions.
• The authors suggest AI may enhance consistency in judicial decision-making.
• The analysis uses Dworkin’s Judge Hercules and Laplace’s Demon as reference points.
• The abstract describes a conceptual and comparative study, not an empirical court test.