Skip to content
Contact
☰
Color scheme: Dark.

Research

What EIL studies, in which spaces and programs, and with what. Observable behavior is studied. No direct access to mental states is claimed.

CENTRAL QUESTION

The questions everything turns around.

Overarching research question

How do new structures, capabilities, cognitive patterns and forms of intelligence emerge, stabilize and transform through interactions within and between complex systems?

Generative research question

Which relational, temporal, cognitive and systemic mechanisms make these transformations possible, and how can they be reconstructed and distinguished from alternative explanations?

Three questions, three levels: the overarching question names what the lab wants to understand, the generative question asks for the mechanisms, and the longitudinal question tests one of them empirically. None replaces the others.

Longitudinal human to AI research question

Can persistent human–AI interaction cause parts of human cognitive practice to become externalized, computationally represented, and recursively reintegrated into the cognitive process that produced them?

The longitudinal question is the specific empirical research question inside this larger program. It asks about one concrete path, the one from human practice through externalization back into the human.

Three levels of investigation
  • Descriptive: What is happening?
  • Explanatory: How and under which conditions did it arise?
  • Experimental: How can the proposed explanation be examined, challenged or tested?

Every research space works on all three levels in its own way. The levels connect the spaces without making them one.

What every statement keeps apart
  • Observed: What is observed?
  • Possible origin: How could it have come about?
  • Evidence: What evidence supports the proposed explanation?

The site keeps these three apart everywhere: an observation is not an explanation, and a plausible explanation is not yet evidence for it.

Axis question, loop and context

Axis question, the roof

This roof belongs to the longitudinal human to AI line. The whole research architecture hangs under the overarching and the generative question above.

How do persistent human–AI interactions produce, externalize and transform cognitive architectures over time?

The axis question is the roof the programs hang under. The main question is its sharpest, testable form. It asks for proof that a part of this architecture really arises through externalization and reintegration.

The decisive empirical question is whether the interaction leaves behind computational structures that later become part of how that person thinks and works.

Externalization is the bridge between ephemeral interaction and durable computational structure. The stronger and more testable claim is that recurring cognitive practices can leave persistent technical traces.

The word “cause” in the question is a target to test. It is not presupposed. Retrospective data can show sequence, association and a plausible mechanism. Stronger claims need prospective comparison.

The externalization loop
  1. Human
  2. Externalization
  3. Artificial function
  4. Effect back on the human

Human cognitive practice, externalized computational structure, reintegration, changed cognitive practice, next computational structure.

First a person works in a particular way. Then that way is put into words and held on record, as a rule, a schema or a role. Then it runs as an artificial function, for example as an agent, a workflow or a review step. Then it acts back on the person's work, and the changed work produces the next structure. Whether the feedback occurs, and how strongly, is the empirical question.

The loop in long form (9 nodes) Concept The nine nodes are at the axis Externalization & Reintegration →

RESEARCH CYCLE

How the results of one phase become the instruments of the next

In the lab, the results of one research phase become the instruments and systems with which the next phase is studied. Papers, methods, systems and experiments are different results of the same program.

  1. Longitudinal human-AI interaction
  2. Observed dynamics
  3. Phenomena and research questions
  4. Models, methods, metrics
  5. Experimental systems and agents
  6. Controlled and naturalistic experiments
  7. New longitudinal evidence
New evidence leads back to interaction and raises new questions.

Observation, emergent phenomena, reconstruction, modeling, instruments, experimentation, new evidence. The cycle is recursive: new evidence changes what is observed next, and instruments built along the way change what can be observed at all. It describes the direction of the work, not a claim that every line of research has already passed every stage.

  • Publicationsdocument phenomena and models (steps 2 to 4).
  • Methodsmake the phenomena measurable (step 4).
  • Systemsapply methods and produce new observations (steps 5 and 6).
  • Programscarry the experiments and hold the questions of the next round (steps 3 and 6).

Research spaces

The eight research spaces form the basic structure of EIL's research. They are of equal rank, and each space investigates its own part of the overarching research question.

  • A studies longer-term interactions between human and AI.
  • B studies human cognition, behavior, self-models and externalization.
  • C studies emergence, self-organization, transformation and complex dynamics.
  • D studies reconstruction, perspectives, relations and possible causal structures.
  • E studies distributed, agentic and collective intelligence.
  • F studies computational representation and modeling.
  • G studies human-machine interfaces and neuroadaptive possibilities.
  • H studies measurement, evidence and scientific infrastructure.

The spaces stay independent. This description connects them; it does not place one under another.

  1. Human–AI Interaction & Co-Evolution

    How do humans, AI systems and their relationship change when interaction continues over long periods?

    A Human–AI Interaction & Co-Evolution

    Human–AI Co-EvolutionLongitudinal Interaction Impact Evaluation

    • Emergence
    • Prediction
    • Adversarial
    • Externalization
    • Longitudinal
    • Memory
    Open space

  2. Human Cognition, Behavior & Externalization

    Which cognitive working practices become visible in long-term human–AI interaction, how do they change, and under which conditions are parts of them externalized?

    B Human Cognition, Behavior & Externalization

    Emergent Cognitive ArchitectureCognitive ExternalizationIntrospection & Self-Analysis ExploratoryConsciousness & Subconscious Processing ExploratoryPerspective Modeling & Cognitive Reconstruction Exploratory

    • Emergence
    • Prediction
    • Externalization
    • Longitudinal
    Open space

  3. Emergence, Adaptation & Complex Dynamics

    When do new, stable and systemically relevant structures arise from repeated interaction that no single element can sufficiently explain?

    C Emergence, Adaptation & Complex Dynamics

    Emergence & Stabilization

    • Emergence
    • Prediction
    • Longitudinal
    • Memory
    Open space

  4. Reconstruction, Perspective & World Models

    How can hidden states, perspectives, relationships, constraints, causal structures and future developments be reconstructed from incomplete evidence?

    D Reconstruction, Perspective & World Models

    Perspective Modeling & Cognitive Reconstruction ExploratoryWorld / Actor / Causal Model Reconstruction

    • Emergence
    • Prediction
    • Adversarial
    • Longitudinal
    Open space

  5. Distributed, Agentic & Collective Intelligence

    How does coordinated intelligence arise when perception, knowledge, decision and action are distributed across several specialized artificial units?

    E Distributed, Agentic & Collective Intelligence

    Distributed Intelligence

    • Emergence
    • Adversarial
    • Externalization
    • Longitudinal
    • Memory
    Open space

  6. Computational Interaction & Cognitive Modeling

    How can interaction trajectories, cognitive working practice, states, transitions and emergent patterns be computationally represented, analyzed and modeled over time?

    F Computational Interaction & Cognitive Modeling

    No dedicated primary program.

    • Emergence
    • Prediction
    • Adversarial
    • Externalization
    • Memory
    Open space

  7. Human–Machine Interfaces & Neuroadaptive Systems

    How do interfaces and adaptive systems influence perception, behavior, cognitive work and human–AI interaction, and how could systems respond in future to measured or inferred human states?

    G Human–Machine Interfaces & Neuroadaptive Systems

    Neuroadaptive Human–AI Systems Planned

    • Emergence
    Open space

  8. Measurement, Evidence & Research Infrastructure

    How can complex longitudinal human–AI systems be observed, measured, reconstructed and studied in a scientifically traceable way?

    H Measurement, Evidence & Research Infrastructure

    No dedicated primary program.

    • Emergence
    • Adversarial
    • Externalization
    • Longitudinal
    • Memory
    Open space

Roadmap of the neuroadaptive direction

Today only the stage built on interaction data and behavioral proxies exists. Everything beyond it is planned and not implemented.

The direction leads from systems that react to explicit input toward systems that adapt to changing states. The loop has eight nodes.

The program Neuroadaptive Human–AI Systems belongs to the space Human–Machine Interfaces & Neuroadaptive Systems. The five stages of the roadmap, the closed loop and the levels of adaptation are there. See the loop in the space Human–Machine Interfaces & Neuroadaptive Systems →

What has come out of the research

EIL studies emergence in technical systems and in human to AI interaction, and equally in human cognition, in long-term patterns of thinking and behavior, and in coupled human to AI systems.

The research asks how new structures arise that did not exist beforehand as a finished unit: cognitive ways of working, self-models, perspectives, strategies, interaction patterns, measures, research instruments, models and technical systems.

Some of these structures were first observed as a recurring pattern, then formalized, and later fed back into the research as an instrument, model or system.

Besides descriptions of emergence, the work of the lab has produced:

  • Research and analysis procedures
  • Measurement systems and metrics
  • Models of human and coupled cognitive processes
  • Procedures for reconstruction, perspective modeling and self-analysis
  • State, memory and evidence structures
  • Specialized multi-agent and intelligence systems

Methods & research instruments →Systems & intelligence architectures →