Patterns for a New Generation: AI and Agents

TL;DR

This work investigates whether design patterns can serve as a computable, shared language for humans and AI agents by encoding patterns in a structured template and driving agent behavior via the Active Inference framework rather than hard-coding rules. It presents a minimal, executable workflow (FuLab) for pattern generation, selection, and use, and demonstrates how pattern maturity can be represented as a precision prior guiding stochastic policy sampling over candidate patterns with softmax over $-G/\tau$. The contributions include a practical pattern-template architecture, an Active Inference–based pattern selector, and two provocations: teaching patterns in education (Camp CryptoBot) and governance/coordination examples (pseudo-virtual C-Suite), illustrating educational benefits, governance challenges, and human–AI collaboration dynamics. The work aims to provide a foundation for pattern-aware AI governance and pedagogy, balancing automation with continued human judgment and oversight, and suggests directions for future verification, scalable education, and responsible AI deployment.

Abstract

Design patterns have been used in various fields of inquiry and endeavour to externalize procedural knowledge in a form that supports human reasoning and coordination. In this paper, we show that contemporary Large Language Model (LLM)-based systems can also read, generate, and reason with design patterns written in a structured template. We describe an experimental workflow in which patterns function as shared priors for action selection, reflection, and revision in hybrid human/agent settings. Drawing on the Active Inference Framework, we illustrate how patterns can guide agent behavior without fully prescribing it. This provides a proof of concept that pattern-capable agents can be created using now-standard software tools. We discuss implications for software development, education, business, and AI governance.

Figures (4)

• Figure 1: Pattern-driven scheme for guiding coding agents
• Figure 2: fucodex run showing aspects of pattern selection and use
• Figure 3: Hands-On, Experiential Learning as Pattern Recognition
• Figure 4: Agent action review (aAR) protocol