Anthromekagogy is the study and design of human-system co-evolution: the reciprocal shaping of human cognition, agency and institutions and the systems people build. It asks whether integration produces durable capability or structural dependency, and treats that as a question to be measured.
Definition
Anthromekagogy is the systematic study and design of feedback architectures between humans and constructed systems, with the objective of increasing durable adaptive intelligence while preserving agency and institutional stability. It is normative in its goal, which is capability expansion, and empirical in its method.
The 2011 formulation treated technology as an extension of human cognition and intelligence amplification as desirable. The 2026 framework keeps that starting point but adds what the earlier version lacked: defined metrics, a clear distinction between empowerment and dependency, engagement with empirical cognitive science, and institutional analysis. It draws on extended and distributed cognition, cybernetics, human-AI collaboration research, institutional economics and complex adaptive systems.
The Human-System Unit
The framework's basic unit of analysis is the Human-System Unit (HSU): a human cognitive agent, a constructed system, a task environment and a two-way feedback channel. Examples include a person using a navigation app, a physician working with a diagnostic model, and a population living with algorithmic media. HSUs scale: organizations and societies are larger HSUs made of smaller ones.
Co-evolution is the reciprocal adaptation that happens inside an HSU over time. People adapt to what a system makes easy, systems adapt to how people behave through the data they collect, and institutions adapt to both.
The co-evolution stack
Anthromekagogy models co-evolution across five interacting layers: cognitive, interface, algorithmic, institutional and ecological. Its vertical propagation principle holds that distortions at lower layers travel upward, while incentives at higher layers reshape optimization below. Co-evolution is multilevel and recursive.
Measuring augmentation and dependency
To keep the idea testable, the framework proposes measurable constructs:
- Augmentation Index (AIx). Compares a person's baseline performance, their performance with the system, and their independent performance after using it, to distinguish durable augmentation from outsourcing and dependency.
- Skill Retention Half-Life. The time for an independently executable skill to decay by half under reliance on a system.
- Automation Bias Rate. How often people accept incorrect system output despite contradicting evidence.
- Epistemic Diversity Index. The variety of information sources across a population, weighted by exposure.
Failure modes and design constraints
The framework identifies five recurring failure modes: automation drift, the skill atrophy loop, metric corruption, epistemic capture, and power concentration. Against them it sets five design constraints: cognitive integrity (systems should preserve or strengthen human competence), calibrated friction (some effort aids learning), feedback transparency (people should understand what a system optimizes, its uncertainty and how it fails), reversibility (people must be able to function without the system), and institutional alignment (incentives should reward long-term capability over short-term engagement).
Ethically, Anthromekagogy rejects both technological determinism and anti-technological retreat. It aligns with the capability approach associated with Amartya Sen: the aim is to expand what people are genuinely able to do and be, while preserving agency and collective resilience.
Applied here: the building as a system
Authentic Intelligence applies the framework to critical infrastructure, where the constructed system is often a building. In a data center, a substation, a water plant or an emergency operations center, the operator is the human agent; the facility, with its envelope, power, access control, automation and models, is the constructed system; the operating condition is the task environment; and alarms, displays, records and automatic actions form the feedback channel.
That framing explains why this site treats physical protection, personnel trust, sensing and materials as one subject. A wall sets how much time an operator has. An alarm design decides whether attention is sharpened or worn down. Automatic transfer logic is an algorithm with delegated authority. Codes and insurance are the institutional layer. Energy supply and materials are the ecological layer. The essays below work through that application:
- Can a Building Be Part of a Human-System Unit?
- Can Machines Guide Operators Without Replacing Their Judgment?
- How Does the Co-Evolution Stack Apply to a Facility?
Open questions
The framework's research program sets out hypotheses to test, including whether calibrated friction improves long-term retention, whether hybrid decision systems outperform both full automation and full manual control, and whether transparency reduces automation bias without costing performance. Its open questions include the right balance of human and algorithmic decisions by domain, and whether reversible augmentation is stable at scale.
Human-system integration is already under way. Whether it builds durable capability or structural dependency will be decided by objective functions, institutional incentives, feedback design and measurement discipline.
References
- Clark, A. and Chalmers, D. (1998). The Extended Mind. Analysis 58(1).
- Wiener, N. (1948). Cybernetics: Or Control and Communication in the Animal and the Machine. Technology Press / John Wiley & Sons.
- Arthur, W. B. (1989). Competing Technologies, Increasing Returns, and Lock-In by Historical Events. Economic Journal 99(394).
- Ong, W. J. (1982). Orality and Literacy. Methuen.
- Dahmani, L. and Bohbot, V. D. (2020). Habitual use of GPS negatively impacts spatial memory during self-guided navigation. Scientific Reports 10.
- Risko, E. F. and Gilbert, S. J. (2016). Cognitive Offloading. Trends in Cognitive Sciences 20(9).
- Kamar, E. (2016). Directions in Hybrid Intelligence: Complementing AI Systems with Human Intelligence. Proceedings of IJCAI 2016.
- Parasuraman, R. and Riley, V. (1997). Humans and Automation: Use, Misuse, Disuse, Abuse. Human Factors 39(2).
- Carr, N. (2010). The Shallows. W. W. Norton.
- Sunstein, C. R. (2001). Republic.com. Princeton University Press.
- Strathern, M. (1997). Improving ratings: audit in the British University system. European Review 5(3).
- Bjork, R. A. (1994). Memory and metamemory considerations in the training of human beings. In Metcalfe, J. and Shimamura, A. (eds.), Metacognition: Knowing about Knowing, MIT Press.
- Sen, A. (1999). Development as Freedom. Oxford University Press.