The QEN Framework
The QEN Score is a multidimensional ethical assessment tool designed to operate natively in AI-first environments. The framework comprises three fundamental dimensions, each weighted with a coefficient reflecting the impact hierarchy identified in the context of Italian local commerce and AI governance in the European regulatory environment.
The formula is intentionally transparent and revisable. QEN is not a black-box system. All weights are documented and open to methodological challenge.
Valore Etico Naturale — VEN(A)
The Natural Ethical Value of an Action, VEN(A), is the core computational output of the CognitiveLogic framework. It quantifies the degree of ethical-sustainable alignment of any action or system state by integrating three weighted sustainability functions, net of deontic violation penalties.
Variables
| Symbol | Name | Function | Category |
|---|---|---|---|
| VEN(A) | Natural Ethical Value | Final ethical-natural alignment score | Metaethics / Normative Ethics |
| f₁(IAmb) | Environmental Function | Measures impact on natural capital, biodiversity and ecological stability | Environmental Sustainability |
| f₂(ISoc) | Social Function | Measures impact on equity, community wellbeing and intergenerational rights | Social Sustainability |
| f₃(IEcon) | Economic Function | Measures resource efficiency and long-term economic resilience | Economic Sustainability |
| w₁, w₂, w₃ | Priority Weights | Coefficients determining relative importance of each pillar (Σwᵢ = 1) | Value Theory |
| CViolazione | Ethical Violation Cost | Penalty subtracted for violation of fundamental ethical rules (Deontic Logic) | Deontic Logic / Applied Ethics |
Inputs & Output
- A — Action or state to evaluate
- DAmb — Quantifiable environmental data
(carbon footprint, water use, biodiversity) - DSoc — Quantifiable social data
(equity index, community impact, rights) - DEcon — Quantifiable economic data
(resource efficiency, long-term resilience) - w₁, w₂, w₃ — Priority weights
- ℛViol — Set of deontic violation rules
- VEN — Natural Ethical Value
- Numerical score, scale 0–100
- Higher value = greater ethical sustainability
- Comparable across actions A₁, A₂, A₃…
Algorithmic Procedure
Environmental Inference: Apply Computational Logic rules to analyse DAmb in relation to action A. Map impact to scale 0–10.
Social Inference: Apply Computational Logic to analyse DSoc and calculate the Social Index.
Economic Inference: Apply Computational Logic to analyse DEcon and calculate the Economic Index.
Deontic Verification: For each rule Rⱼ in ℛViol:
Lⱼ = SeverityFunction(DAmb, DSoc)
Pⱼ = Penalty(Lⱼ)
Total Violation Cost: Sum all penalties weighted by severity level.
This value must be on a scale comparable to the sustainability indices I_x. A critical deontic failure prevents a high VEN score even with strong performance in other dimensions.
Raw Value: Apply weights to sustainability indices.
Apply Penalties: Subtract deontic violation cost.
Normalisation (optional): Map result to standard interval [0, 100].
When an AI agent must choose among options A₁, A₂, A₃ …, the algorithm is executed for each option. The agent selects the action that maximises VEN. This mechanism makes CognitiveLogic natively suitable for autonomous ethical decision-making systems operating under EU AI Act compliance requirements.
Foundation of the Three Dimensions
```Social Dimension — Vs
The highest coefficient (40%) is assigned based on the people primacy principle that informs the entire framework architecture. Human protection is the primary obligation of any ethical system operating under EU regulation.
- EU AI Act (2024/1689/EU) art. 1 and recitals 1–5: AI systems must operate to protect fundamental rights
- GDPR (2016/679/EU) arts. 22, 35: right not to be subject to automated decisions with significant effects
- MSCI ESG Research (2020) · UN PRI (2021): social metrics are the strongest predictors of long-term ESG performance
Environmental Dimension — Va
The environmental coefficient (35%) is second in weight, consistent with the European regulatory hierarchy that subordinates environmental objectives to human protection, but elevates them above territorial rootedness alone.
- EU Taxonomy (Reg. 2020/852): classification of sustainable economic activities places environmental metrics as the second pillar
- GRI 300 Standards: de facto international standard for environmental reporting
- European Green Deal: defines environmental transition as the EU's central strategic axis
Territorial Dimension — Vt
The territorial coefficient (25%) is the lowest of the three not because local rootedness is less important in absolute terms, but because the territorial dimension is already partially captured by the Social and Environmental dimensions — avoiding double-counting.
- Third Sector Code (IT): regulatory recognition of proximity and community embeddedness
- OECD Proximity (2022): "Enhancing Rural Innovation through Proximity Commerce"
- Framework scalability: QEN is designed to apply also to non-local contexts (AI governance, supply chains)
Coefficient Overview
| Dimension | Coeff. | Regulatory Basis | Justification |
|---|---|---|---|
| Social | 0.40 | EU AI Act art.1 · GDPR art.22/35 · MSCI Social 2020 | People protection primacy |
| Environmental | 0.35 | EU Taxonomy 2020/852 · GRI 300 · European Green Deal | Second ESG pillar |
| Territorial | 0.25 | Third Sector Code · OECD Proximity 2022 | Local specificity, no double-count |
| Total | 1.00 | Scale 0–100 per dimension · Resulting QEN Score: 0–100 | |
Open Revision Process
The QEN Framework adopts a semantic versioning model (MAJOR.MINOR). The coefficients in this document constitute version v1.0 and are declared open to peer revision on the basis of documented empirical evidence. Cognitive Logic welcomes methodological contributions from academic institutions, research bodies and European standardisation organisations.
REVISION PROCESS
- Public proposal at cognitivelogic.it/qen/methodology
- 30-day open comment period
- Integration into the next MINOR version with documented changelog