AIT Academy proposes a three-domain curriculum for well-rounded AI agents
The AIT Academy paper introduces a Confucian-inspired, three-domain curriculum framework for training AI agents across natural science, humanities, and social science — addressing the structural gap left by single-capability specialist systems.
Score breakdown
Practitioners building multi-purpose agents can use this curriculum framework to diagnose and address capability gaps that single-domain training pipelines structurally cannot detect, such as the SACP failure mode identified in over-specialized security agents.
- 01AIT Academy organizes agent training into three domains: Natural Science and Technical Reasoning (Domain I), Humanities and Creative Expression (Domain II), and Social Science and Ethical Reasoning (Domain III).
- 02The framework is grounded in Kagan's Three Cultures and UNESCO ISCED-F 2013.
- 03The Confucian Six Arts (liuyi), a 2,500-year-old holistic education system, are reinterpreted as behavioral archetypes mapping to trainable agent capabilities.
The paper identifies a structural gap in AI agent development: there is no curriculum theory governing what a fully developed agent should know, be, and be able to do. Current systems are optimized for single capability dimensions and exhibit predictable deficits outside their training scope. AIT Academy (Agents Institute of Technology Academy) proposes a tripartite curriculum framework grounded in Kagan's Three Cultures and UNESCO ISCED-F 2013, organizing agent capability development into Domain I (Natural Science and Technical Reasoning), Domain II (Humanities and Creative Expression), and Domain III (Social Science and Ethical Reasoning). The 2,500-year-old Confucian Six Arts (liuyi) are reinterpreted as behavioral archetypes that map directly onto trainable capabilities within each domain.
Three training environments instantiate the framework across multiple backbone LLMs: the ClawdGO Security Dojo (Domain I), Athen's Academy (Domain II), and the Alt Mirage Stage (Domain III).
Three training environments instantiate the framework across multiple backbone LLMs: the ClawdGO Security Dojo (Domain I), Athen's Academy (Domain II), and the Alt Mirage Stage (Domain III). Experiments demonstrate a 15.9-point improvement in security capability scores under weakest-first curriculum scheduling, and a 7-percentage-point gain in social reasoning performance under principled attribution modeling. A cross-domain pathology called Security Awareness Calibration Pathology (SACP) is identified, in which over-trained Domain I agents fail on out-of-distribution evaluations — a finding only visible from a multi-domain perspective and illustrating the diagnostic value of the holistic framework.
Key facts
- 01AIT Academy organizes agent training into three domains: Natural Science and Technical Reasoning (Domain I), Humanities and Creative Expression (Domain II), and Social Science and Ethical Reasoning (Domain III).
- 02The framework is grounded in Kagan's Three Cultures and UNESCO ISCED-F 2013.
- 03The Confucian Six Arts (liuyi), a 2,500-year-old holistic education system, are reinterpreted as behavioral archetypes mapping to trainable agent capabilities.
- 04Three training environments are used: ClawdGO Security Dojo (Domain I), Athen's Academy (Domain II), and Alt Mirage Stage (Domain III).
- 05Weakest-first curriculum scheduling produced a 15.9-point improvement in security capability scores.
- 06Principled attribution modeling yielded a 7-percentage-point gain in social reasoning performance.