The concept of a “nimbusian” legal system driven by personalized, real-time directives is intriguing. I was reminded of the “silver box” and “gold box” legal solutions model:
Silver box problems are characterised by being (relatively):
- rules based;
- stable and predictable;
- repeatable; and
- scalable.
Silver box problems are therefore amenable to:
- the application of collaborative problem solving to create a model;
- the application of a process and lean thinking;
- the application of data driven knowledge and insights;
- the application of technology; and
- a train, maintain, sustain approach.
Gold box problems are characterised by being:
- complex, multifaceted and ambiguous;
- unpredictable and uncertain;
- rapidly changing or chaotically decaying; and
- impacted by irrationality, emotion, dishonesty and bias.
Gold box problems are therefore amenable to:
- the application of collaborative problem solving to solve each problem;
- harnessing diverse thinking and sources of insight; and
- agility and responsiveness.
The nimbusian system seems well-suited to address silver box problems. Strong protocols and automation could be applied to streamline processes like contract enforcement, regulatory compliance, and possibly aspects of intellectual property.
However, the complexities of gold box problems require a more nuanced approach. While the nimbusian system’s personalized directives could be beneficial in these areas, the inherent ambiguity and reliance on human judgment in these cases suggest the need for weak protocols that allow for flexibility and discretion.
Although I agree that protocols and technology have the potential to improve access to justice, the nimbusian system’s reliance on real-time data and algorithmic decision-making raises questions about fairness and accountability. In this context, it’s important to recognize the limitations of the underlying reasoning mode that drives much of current AI. Machine learning is primarily based on inductive reasoning. While powerful, inductive reasoning yields inferences that are likely, rather than certain. In a legal context, relying solely on probabilistic conclusions derived from machine learning models could be problematic, particularly in gold box problems, which are complex, ambiguous, and often influenced by irrational factors.
A hybrid approach combining strong and weak protocols, automation, and human judgment could be the most effective way forward.