Artificial Agency Program: Curiosity, compression, and communication in agents

2026-02-27

AI Summary

This work investigates how to construct embodied agents driven by curiosity—formalized as learning progress—as an intrinsic motivation under constraints of physical and computational resources, and how to effectively integrate such agents into human–tool–environment systems. To this end, the authors propose a unified framework that jointly incorporates predictive compression, intrinsic motivation, empowerment-based control, and language-based self-communication. Grounded in the principles of the information bottleneck and bounded rationality, the framework dynamically allocates limited resources across perception, action, and reasoning. The project’s contributions include a testbed featuring an explicit cost model and staged experimental design, which bridges intrinsic motivation theory with real-world embedded AI systems through multimodal tokenization experiments informed by information theory, thermodynamics, and modern reasoning architectures.

Technology Category

Multiagent Systems: Mechanism DesignCognitive Modeling & Cognitive Systems: Agent ArchitecturesIntelligent Robots: Embodied AI

Application Category

Responsible Web: Machine-in-the-loop, human agency and autonomyEconomics, Online Markets and Human Computation: Incentives in network design for Web infrastructures and ecosystemsSemantics and Knowledge: Data modeling to support human-machine intelligence, including LLMs agents, intelligent system behavior, explanations, and user-friendly interactions

Abstract

This paper presents the Artificial Agency Program (AAP), a position and research agenda for building AI systems as reality embedded, resource-bounded agents whose development is driven by curiosity-as-learning-progress under physical and computational constraints. The central thesis is that AI is most useful when treated as part of an extended human--tool system that increases sensing, understanding, and actuation capability while reducing friction at the interface between people, tools, and environments. The agenda unifies predictive compression, intrinsic motivation, empowerment and control, interface quality (unification), and language/self-communication as selective information bottlenecks. We formulate these ideas as a falsifiable program with explicit costs, staged experiments, and a concrete multimodal tokenized testbed in which an agent allocates limited budget among observation, action, and deliberation. The aim is to provide a conceptual and experimental framework that connects intrinsic motivation, information theory, thermodynamics, bounded rationality, and modern reasoning systems

Problem

• Artificial Agency
• Curiosity
• Resource-bounded Agents
• Human-Tool Systems
• Information Bottlenecks

Innovation

• Artificial Agency
• intrinsic motivation
• predictive compression
• bounded rationality
• multimodal tokenized testbed