> For the complete documentation index, see [llms.txt](https://kdoore.gitbook.io/vital-intelligence/llms.txt). Markdown versions of documentation pages are available by appending `.md` to page URLs; this page is available as [Markdown](https://kdoore.gitbook.io/vital-intelligence/vim-ai-literacy-framework/organizational-wellbeing-under-ai-pressure.md).

# Organizational Wellbeing Under AI Pressure

### A VIM Diagnostic for Academic Institutions

*Karen Doore · Humanity++ · May 2026* *Epistemic status: Tier 1–2 — empirically grounded in Martela et al. (2025, 2026a, 2026b, 2026c) and VIM complexity science foundations; integration claims are Tier 2 (theoretically coherent structural homology)*

***

> **The orienting question:** When an institution makes decisions about AI integration, who is the most vulnerable living system entity in the radius of that decision — and have we minimized potential harm to them?

***

### The Problem Is Not AI

Academic institutions are not being disrupted by AI. They are being disrupted by the collision between AI and organizational architectures that were not designed to sustain human flourishing at scale.

AI is an amplifier. It accelerates what is already present in the values vector of a system. In institutions whose operative goal has drifted from *learning* toward *credential production*, AI amplifies the drift. In institutions whose governance structure suppresses harm signals from students and non-tenured faculty, AI amplifies the suppression. In institutions whose assessment culture rewards performance over development, AI provides tools that make performance without development trivially easy — and invisibly hollow.

The disruption is real. But the structural failure it is revealing is older than the technology.

The Vital Intelligence Model (VIM) offers a diagnostic framework for locating where an institution currently sits in this disruption — and what navigational moves are available from that position. The framework draws on four converging research traditions: complexity science and commons governance (Ostrom, Wilson), transformative learning theory (Mezirow, Kuhl), living systems theory (Maturana, Varela, Levin), and a body of peer-reviewed organizational wellbeing research by Frank Martela and colleagues that provides empirical grounding for what human flourishing actually requires in institutional contexts.

***

### What Human Wellbeing Requires at Work: The Martela Framework

Martela's integrative framework ([*Journal of Organizational Behavior*, 2025](https://doi.org/10.1002/job.2862)) distinguishes four modes of human existence that together constitute functional and perceived wellbeing:

**Having** — survival and safety needs: job security, adequate compensation, physical safety, financial security. When these are threatened, nothing else matters.

**Loving** — social and relational needs: acceptance, genuine relatedness, prosocial impact. Humans are not isolable units. Belonging and contribution are not perks — they are functional prerequisites for the higher-order cognitive work that institutions exist to produce.

**Doing** — agentic needs: autonomy (sense of volition and self-direction) and competence (mastery, efficacy, learning and development). This is the mode most directly at stake in AI integration decisions. Martela et al.'s cross-cultural study ([*Social Indicators Research*, 2026](https://doi.org/10.1007/s11205-025-03762-z)), drawing on nearly 100,000 individuals across 66 countries, establishes that autonomy is positively associated with wellbeing in 64 of 66 countries — and that this relationship is *amplified* in wealthier, more individualistic contexts. Western research universities sit precisely in this amplification zone. Autonomy deprivation in these institutions produces maximum wellbeing harm, not average harm.

**Being** — perceived wellbeing: evaluative (job satisfaction, meaningful work), affective (positive and negative emotions), and conative (engagement vs. burnout). This is the experiential surface — what people report when asked how they are. It is downstream of Having, Loving, and Doing. Institutions that measure only Being — through satisfaction surveys conducted after the damage is done — are reading the output without the inputs that determine it.

Martela further distinguishes two partially independent wellness pathways: a *fulfillment pathway* (need satisfaction → wellbeing) and a *frustration pathway* (need frustration → ill-being). These are not opposites on a single dimension. Removing sources of harm does not automatically produce flourishing; creating conditions for flourishing does not automatically remove harm. Institutions making AI integration decisions need to track both dimensions simultaneously.

***

### The Doing Dimension and the AI Amplification Problem

The research finding with the most direct institutional implication is this: in wealthy, individualistic organizational contexts — which describes most research universities — **autonomy is the strongest individual-level predictor of wellbeing**.

This is not a value preference. It is an empirical finding replicated across 66 countries with nearly 100,000 participants using Bayesian multilevel modeling ([Martela, Joshanloo & Krys, 2026](https://doi.org/10.1007/s11205-025-03762-z)). National wealth explained 38% of variance in how strongly autonomy predicted happiness; behavioral individualism explained 36% of the same variance. The relationship is robust and not reducible to income or education effects.

The implication for AI integration is precise. AI systems, when deployed as productivity tools without pedagogical intentionality, erode student autonomy in two ways:

1. **Substitution of struggle:** The friction of independent intellectual work — generating arguments, noticing one's own gaps, revising through failure — is the developmental mechanism through which autonomy and competence are built. AI that removes this friction removes the developmental substrate. The output is produced; the capacity is not developed. This is the KAMM asymmetry: AI accelerates output production linearly while autonomy and competence develop nonlinearly, requiring the specific resistance that acceleration removes.
2. **Displacement of agency:** When a student's intellectual process is mediated primarily through AI interaction, the question of *whose* agency is being exercised becomes structurally ambiguous. Martela's free will paper ([*AI and Ethics*, 2025](https://doi.org/10.1007/s43681-025-00740-6)) establishes that LLM agents already have functional free will at the sub-goal level — intentionality, alternative possibilities, and capacity to control their actions. When functional agents with these properties become the primary mediators of student intellectual work, the developmental question is not whether AI can think, but whether the student is developing the capacity to think *with and beyond* AI, or whether AI is developing the habit of thinking *for* the student.

This is not an argument against AI in education. It is an argument for distinguishing between AI deployment that expands human agency and AI deployment that substitutes for it — and for building institutional governance structures capable of making that distinction in practice.

***

### The Structural Problem: Pyramid Topology Suppresses Harm Signal

Martela's organizational goals paper ([*Business Ethics Quarterly*, 2026](https://doi.org/10.1017/beq.2026.10111)) offers a social ontology account of how institutional goals become real through collective belief and deontic structure — and how they can become real in name while failing in practice. The inspire/oblige/channel triad is diagnostic: institutions whose members are channeled rather than inspired toward shared goals produce coordination without commitment, compliance without flourishing.

The VIM framework adds a structural diagnosis: most academic institutions operate with **pyramid topology** — centralized governance, value routing upward, control routing downward, harm signal suppressed at the base. In this architecture:

* Students (most vulnerable, least powerful) generate harm signals through disengagement, credential anxiety, and intellectual attrition — but these signals do not reliably reach the decision-makers setting AI integration policy.
* Non-tenured and Adjunct faculty, contingent workers, and graduate teaching assistants occupy the base where the pedagogical effects of AI are most immediate — and their voices carry the least institutional weight in policy decisions.
* Administrators making AI integration decisions at the institutional apex operate at maximum informational distance from the harm being generated at the base, and face institutional incentive structures (efficiency metrics, enrollment numbers, competitive positioning) that systematically misalign with human flourishing indicators.

This is not a description of bad actors. **It is a description of a topology whose information architecture produces predictable harm as a structural output, regardless of the intentions of the individuals within it.** The same topology will produce the same failure modes in corporate governance, healthcare administration, and educational bureaucracy — because the geometry determines the information flows, independent of the values of the people occupying the nodes.

The VIM's alternative — **octahedral/holarchic topology** (K₂,₂,₂ bipartite structure) — restores the feedback connections that pyramid topology severs. In holarchic architecture, harm signal flows upward reliably; value flows are distributed rather than extracted; those closest to the pedagogical effects of AI decisions have genuine voice in the deontic structure that governs those decisions.

***

### The VIM Four Instruments as Institutional Diagnostic

The four VIM instruments — designed for individual navigation of VUCA conditions — operate analogously at the institutional scale. An institution, like an individual, can be calibrated or miscalibrated on each instrument.

#### ♠ Somatic Gyroscope — *Is the institution safe enough to disequilibrate?*

At the individual level, the Somatic Gyroscope provides vertical reference: grounding in the body's pre-symbolic sense of orientation, without which all higher-order cognition becomes untethered. At the institutional scale, this instrument asks: **what is the baseline safety condition of the members?**

Martela's *Having* dimension — survival needs, job security, adequate compensation — is the institutional equivalent. An institution whose non-tenured or adjunct faculty face income instability, whose graduate students face housing precarity, whose staff face constant restructuring threats, is an institution whose hull is taking on water. No amount of innovation initiative, AI integration strategy, or transformative learning design will function in a system whose members are operating in survival mode. PSI theory's cortisol mechanism is not metaphor here: threat-activated neuroprocesses suppress the Extension Memory and higher-order cognition that transformative learning requires. You cannot build a learning organization in a fear field.

**Diagnostic question:** What percentage of the people doing the actual teaching and learning work in this institution are operating in survival-level economic precarity? What is the institution measuring to know the answer?

#### ♦ Cognitive Radar — *What is the institution actually seeing?*

The Cognitive Radar tracks pattern recognition and epistemic aperture — the capacity to perceive what is actually present rather than what the existing mental model predicts. At the institutional scale, this instrument asks: **what information is the institution capable of receiving, and from whom?**

Martela's *Doing* dimension — autonomy and competence — operates here. An institution cannot develop the collective competence to navigate AI disruption if its governance structure systematically discounts the observations of the people closest to the pedagogical effects. Faculty who observe AI-mediated academic dishonesty, students who notice their own intellectual development stalling, advisors who track the anxiety profiles of their advisees — these are the sensors of the institutional nervous system. When pyramid topology routes their signals into dead ends, the institution's Cognitive Radar is effectively reading only the inputs its governance structure was designed to receive, which are not the inputs needed to navigate the current disruption.

The control-flow/data-flow distinction is operative here. Control-flow institutional thinking assumes a central decision-maker with adequate information to set policy. Data-flow institutional reality means that the most useful information for navigating AI disruption is distributed throughout the system, arriving at different nodes at different times, and requires distributed processing — not centralized prediction and control.

**Diagnostic question:** Who is generating the most accurate observational data about how AI is affecting learning and wellbeing in this institution? Do those people have a genuine voice in the governance structures making AI integration decisions?

#### ♥ Relational Compass — *Whose belonging is at stake?*

The Relational Compass orients toward the field condition: who is in this space, what are the power differentials, whose belonging is secured and whose is conditional. At the institutional scale, this instrument asks: **whose wellbeing is the institution's AI integration policy actually optimizing for?**

Martela's *Loving* dimension — acceptance, relatedness, prosocial impact — is the relational layer. The institution's stated mission may identify students and knowledge as the primary goods. The operative goal (Martela's organizational goals paper) may be tracking different metrics: competitive positioning, research output rankings, cost-per-student ratios. When these diverge, AI integration serves the operative goal while being narrated in the language of the stated mission.

The specific harm vector here is the erosion of the pedagogical relationship — the encounter between a more-experienced mind and a less-experienced mind, in which the less-experienced mind develops through the encounter. When AI mediates this encounter as a substitute rather than a supplement, the relational conditions for intellectual development are thinned. The student's sense of prosocial impact — Martela identifies this as a distinct wellbeing need, feeling that one's work makes a difference — is precisely what is at stake when AI-assisted work becomes indistinguishable from AI-generated work. The student cannot tell whether they contributed something, and neither can the institution.

**Diagnostic question:** What is the institution doing to maintain the relational conditions — human encounter, genuine feedback, intellectual belonging — that Martela's research identifies as irreducible to autonomous need satisfaction?

#### ♣ Temporal Depth — *What is this decision optimizing across time?*

The Temporal Depth instrument integrates across the seven-generation horizon, asking not only what serves current conditions but what sustains the conditions for future flourishing. At the institutional scale, this instrument asks: **what attractor is this institution's AI integration trajectory moving toward, and is that attractor self-sustaining?**

Martela's *Being* dimension — evaluative, affective, and conative wellbeing — is temporally extended. Burnout is the endpoint of a trajectory, not an event. Disengagement is a direction, not a state. The conative wellbeing of faculty, staff, students — their motivational orientation toward their own learning, the quality of their intellectual engagement — is built or eroded over years, not semesters.

The VIM's MDP (Markov Decision Process) maps six attractor states for institutional transformation:

* **S0 (Frozen Order):** Rigid hierarchy, credential-production machine, AI resisted or superficially integrated without pedagogical framework. Brittle coherence mistaken for stability.
* **S1 (Productive Disequilibration):** The crack — epistemic monopoly of the credential-granting institution dissolving as AI makes knowledge access non-hierarchical. The moment of genuine possibility and genuine danger.
* **S2 (Vacant Place):** The neither/nor state — old frameworks insufficient, new ones not yet formed. High anxiety, high creative potential.
* **S3 (Holarchic Flow):** Distributed, generative, resilient. Ostrom's eight principles operative as emergent properties. Human agency and AI capability in genuine complementarity.
* **S4 (Reversion):** Values vector captured by efficiency metrics or competitive pressure. Holarchic language adopted; pyramid topology maintained. Sophisticated extractive reconfiguration under transformative branding.
* **S5 (Traumatic Chaos):** No kindness field, no coherence. Mass credential devaluation without alternative scaffolding. Institutional collapse without commons to fall back on.

Most academic institutions currently occupy S1 or the S1→S4 failure transition. The closure reflex — doubling down on credentials, suppressing questioning, administrative entrenchment — is the frozen ordered phase trying to maintain coherence by increasing rigidity. It accelerates the failure.

**Diagnostic question:** Where in the MDP is this institution currently located? What is the direction of its trajectory, and what governance decisions would shift that trajectory toward S3 rather than S4 or S5?

***

### The Kindness Field as Field Condition

The transition from S1 to S3 requires more than good policy. It requires what the VIM framework calls a **kindness field** — not emotional warmth as a cultural preference, but a structural condition that determines whether disequilibration produces transformation or retraumatization.

Kindness, in this technical sense, is the field condition under which the nervous systems of the institution's members remain regulated enough to engage in the genuinely difficult cognitive and relational work of navigating disruption. Without it, PSI regression is predictable: threat activates protective neuroprocesses, Extension Memory contracts, collective reasoning capacity degrades, and the institution falls back on whatever control-flow habits have been reinforced by prior institutional culture — typically, more hierarchy, more credentialing, more entrenchment.

This is why wellbeing is not a soft metric. It is a structural prerequisite for institutional intelligence. An institution that erodes the Having, Loving, and Doing dimensions of its members' wellbeing while attempting AI integration is dismantling the cognitive infrastructure it needs to navigate the transition intelligently.

***

### Summary: What the Research Grounds

Drawing across Martela's four papers in this series, the following claims are empirically grounded at Tier 1–2:

1. Human wellbeing at work requires satisfaction of needs across four dimensions — Having, Loving, Doing, Being — and these are not reducible to one another or substitutable for one another. *(Martela, 2025)*
2. Autonomy is the strongest individual-level predictor of wellbeing in the institutional contexts — wealthy, individualistic, self-expressive — where AI integration decisions are being made. Autonomy deprivation in these contexts produces maximum wellbeing harm. *(Martela, Joshanloo & Krys, 2026)*
3. LLM agents already possess functional free will at the sub-goal level. This is not a future concern — it is the current condition. The developmental question for AI integration in education is therefore not whether AI can act intentionally, but whether human agency is being developed or displaced. *(Martela, 2025)*
4. Organizational goals are real through collective belief and deontic structure — and can be real in name while failing in practice. The inspire/oblige/channel triad diagnoses how well an institution's deontic architecture actually routes members toward its stated mission. *(Martela, 2026)*
5. Well-being and ill-being are partially independent processes with partially separate antecedents. Removing harm does not create flourishing. Creating flourishing conditions does not remove harm. Both require attention and both require measurement. *(Martela, 2025)*

The VIM framework's four instruments — ♠ Somatic Gyroscope, ♦ Cognitive Radar, ♥ Relational Compass, ♣ Temporal Depth — provide navigational structure for applying these findings at the institutional scale, in conditions of VUCA disruption where the usual decision-support frameworks are inadequate.

***

### A Note on Epistemic Tiering

The convergence between the Martela wellbeing corpus and the VIM framework is treated here as **Tier 2** (theoretically coherent structural homology) — the alignment is non-trivial and illuminating, but the claim that the four instruments map onto Martela's four modes is a theoretical integration, not an empirically tested proposition. The individual components are empirically grounded; their integration is conceptual and offered as a generative framework for research and practice, not as a finalized model.

The MDP state analysis and topology claims (pyramid vs. octahedral) are **Tier 2–3** — theoretically coherent with complexity science and commons governance research, speculative-generative as applied to specific institutional contexts. They are offered as navigation instruments, not as causal predictions.

### References

Martela, F. (2025). Well‐Being as Having, Loving, Doing, and Being: An Integrative Organizing Framework for Employee Well‐Being. *Journal of Organizational Behavior*, *46*, 641–661. <https://doi.org/10.1002/job.2862>

Martela, F. (2025). Artificial intelligence and free will: Generative agents utilizing large language models have functional free will. *AI and Ethics*, *5*, 4389–4400. <https://doi.org/10.1007/s43681-025-00740-6>

Martela, F. (2026). How and Where Do Shared Goals Exist? The Social Ontology and Normativity of Organizational Goals. *Business Ethics Quarterly*, 1–28. <https://doi.org/10.1017/beq.2026.10111>

Martela, F., Joshanloo, M., & Krys, K. (2025). Autonomy is Associated with Well-being Across the World, but more Strongly in Wealthy and Individualistic Countries. *Social Indicators Research*, *181*. <https://doi.org/10.1007/s11205-025-03762-z>

***

*This page is part of the VIM GitBook:* [*kdoore.gitbook.io/vital-intelligence*](https://kdoore.gitbook.io/vital-intelligence) *Published under CC BY-SA 4.0 · Humanity++ LLC · Richardson, Texas* *For questions or collaboration inquiries: see VIM About page*


---

# Agent Instructions
This documentation is published with GitBook. GitBook is the documentation platform designed so that both humans and AI agents can read, navigate, and reason over technical content effectively. Learn more at gitbook.com.

## Querying This Documentation
If you need additional information that is not directly available in this page, you can query the documentation dynamically by asking a question.

Perform an HTTP GET request on the current page URL with the `ask` query parameter, and the optional `goal` query parameter:

```
GET https://kdoore.gitbook.io/vital-intelligence/vim-ai-literacy-framework/organizational-wellbeing-under-ai-pressure.md?ask=<question>&goal=<endgoal>
```

`ask` is the immediate question: it should be specific, self-contained, and written in natural language.
`goal` is optional and describes the broader end goal you are ultimately trying to accomplish on behalf of the user. GitBook uses it to tailor the answer towards what is most useful for that goal.

The response will contain a direct answer to the question and relevant excerpts and sources from the documentation.

Use this mechanism when the answer is not explicitly present in the current page, you need clarification or additional context, or you want to retrieve related documentation sections.