WFF provides the first empirically proven Constitutional AI Infrastructure for democratic governance. While standard LLMs fail 100% of the time in complex resource governance tasks (defaulting to extractive logic), our neurosymbolic architecture has demonstrated 100% constitutional compliance and 78% self-healing rates across 60 independent trials.

We are seeking $75k to transition this technology from a validated prototype to an open-source standard, deploying it to 10 waiting Community Land Trusts and partnering with the European Urban Initiative to prevent the capture of democratic institutions by misaligned AI.

(Note: This project is ready for immediate execution. Codebase and experimental data are available for review.)

What are this project's goals? How will you achieve them?

Primary Goal: Establish WFF as the open-source standard for AI-assisted governance before extractive, black-box models lock in.

The Problem: AI agents are already entering governance (urban planning, resource allocation). Standard models (GPT-4, Claude) are trained on extractive data. In our control trials, these models produced illegal or extractive governance proposals 100% of the time when left unchecked. There is currently no safety infrastructure to prevent this.

Our Solution: We built the Wisdom Forcing Function (WFF). It is not a prompt; it is a "Verified Dialectical Kernel"—a digital constitution that wraps the LLM. It mathematically enforces regenerative principles (e.g., Circular Metabolism, Distributed Authority).

• Binary Success: 0% viability without WFF →→ 100% viability with WFF.

• Antifragility: The system detects its own violations and self-corrects (78% recovery rate).

How We'll Achieve This:

1. Scientific Validation (Months 1-3): Publish our n=60 experimental data to establish the "Constitutional Standard."

2. Open Source Release (Months 3-6): Release the TypeScript/Python codebase and "Practitioner Toolkit" to prevent vendor lock-in.

3. Deployment (Months 4-9): Deploy to our pipeline of 10 Community Land Trusts and European Urban Initiative pilots.

How will this funding be used?

This funding buys Engineering Velocity and Infrastructure, not just research time.

Budget Allocation ($75k):

• Platform Development ($25k): Refactoring the current prototype into a deployable, open-source package (GUI, API, Documentation).

• Community Deployment ($20k): Technical implementation costs for the 10 pilot communities (Server costs, configuration, on-site support).

• Scientific Validation ($18k): Data analysis, replication studies, and open-access publication fees to cement the standard.

• Training & Dissemination ($10k): Training 50 practitioners to use the system.

• Operations ($2k): Project management.

Minimum Funding ($25k): We focus solely on the Open Source Release and Scientific Proof. This ensures the code exists as a public good, even if we cannot fund the direct deployments.

Who is on your team? What's your track record?

Principal Investigator: Carlos Arleo (PhD Candidate, Newcastle University)

• Unique Fit: 15 years in participatory governance + 12 months developing the WFF neurosymbolic architecture.

• Track Record:

  • Built and deployed the WFF architecture solo.

  • Achieved independent mathematical convergence with Norman Sieroka’s "Universal Governed Order" meta-law.

  • Already deployed in 5 pilot communities with zero instances of extractive capture.

Collaborators & Network:

• Active engagement with European Urban Initiative networks.

• Pipeline of UK Community Land Trusts requesting the technology.

What are the most likely causes and outcomes if this project fails?

Failure Scenarios:

1. Adoption Friction: Communities find the system too technical.

   • Mitigation: The $75k budget prioritizes a "Practitioner Toolkit" (GUI) specifically to solve this.

2. Academic Gatekeeping: Top-tier journals reject the novel "neurosymbolic" approach.

   • Mitigation: We release as Open Source immediately. The code works regardless of the paper's status.

3. Complexity Ceiling: Scaling to "Level 6 Autopoiesis" (Self-Writing Code) proves too computationally expensive.

   • Mitigation: The current "Level 5.8" system (Self-Healing) is already fully functional and sufficient for most governance tasks.

Outcome if project fails:
If we do not fund this, the default future is Extractive AI Governance. Commercial, black-box models will be sold to city councils, optimizing for "efficiency" rather than "democracy." WFF is the only existing counter-measure.

How much money have you raised in the last 12 months?

Why this request?
We have transitioned from "Academic Theory" to "Working Technology."

• We have the code.

• We have the data (n=60 trials).

• We have the demand (European Urban Initiative).

We are seeking Manifund support to bridge the gap between Prototype and Public Infrastructure.

Constitutional Physics: Autopoiesis and Metastability in a Self-Correcting Governance AI https://doi.org/10.5281/zenodo.17604231