You're pledging to donate if the project hits its minimum goal and gets approved. If not, your funds will be returned.
FASO is a public-interest AI safety project developing a neutrality-preserving, reproducible observatory for post-deployment AI drift and risk conditions. The project has moved from design architecture into working prototype form, with internally tested core pipeline components. Newcastle University has agreed to begin formalising a validation pathway, likely involving controlled sandbox testing and open-source AI systems.
This project seeks USD200,000 to fund a validation bridge: controlled external testing of whether the FASO prototype can preserve evidence, route observations through bounded validation and questioning, support replay and audit, resist premature claims, and produce publication-safe artefacts only where sufficiency conditions are met. The purpose is not to claim that FASO is validated already, but to test whether it behaves as a bounded, replayable and non-overclaiming instrument under controlled conditions.
The project will prepare and run the first external validation phase for the FASO working prototype. It will support Newcastle-linked sandbox validation design, controlled testing against open-source or sandboxed AI systems, technical documentation, IP/legal protection, prototype hardening, evidence-pack preparation, and an externally credible validation report.
The validation will test repeatability, sensitivity, specificity and auditability. It will examine whether FASO can distinguish stable behaviour, controlled variation and known-change conditions without over claiming drift or turning visual outputs into unsupported conclusions.
The funding will be used for a focused FASO validation bridge, not for full institutional scale-up.
It will support:
1. Newcastle-linked sandbox validation design and coordination.
2. Controlled testing against open-source AI systems or sandboxed test models.
3. Technical documentation and evidence-pack preparation.
4. Prototype hardening required for external review.
5. IP/legal protection and controlled disclosure support.
6. Independent review and validation-report preparation.
7. Project management and governance support to keep the validation phase bounded, auditable and non-influenceable.
The aim is to test whether the working FASO prototype can preserve evidence, route observations through bounded validation and questioning, support replay and audit, resist premature claims, and produce publication-safe artefacts only where sufficiency conditions are met.
The current core team is:
Nigel Wilson — Founder and design lead. Nigel has led the FASO architecture, doctrine, prototype build and testing process. His background includes senior regional business leadership and complex systems/process work, including resolving large multidisciplinary operational system failures.
Simon Hartley — Project lead / project director support. Simon is supporting the transition from founder-led prototype work into controlled project planning, funding, and external validation.
Newcastle University route — Newcastle University has agreed to begin formalising a validation pathway with us, likely around controlled sandbox testing and possible use of open-source AI systems. This is not yet a completed validation partnership, but it is the intended route for external scrutiny.
FASO has moved from design architecture into working prototype form, with internally tested core pipeline components. The project’s track record so far is founder-led architecture, controlled doctrine development, iterative software build, internal dry-run testing and preparation for external validation.
The most likely failure causes are:
1. The prototype needs more hardening before external testing.
2. Newcastle’s formal process takes longer than expected.
3. The open-source/sandbox validation route proves harder to design than expected.
4. The validation tests show that parts of the evidence chain are insufficient, unclear, or not yet robust enough.
5. Funding is not sufficient to support the necessary documentation, legal/IP protection, technical hardening and review process.
The main outcome of failure would not be unsafe deployment, because FASO is not being deployed live through this project. The outcome would be an explicit finding that the prototype is not yet ready for validation acceptance, followed by redesign, narrowing of scope, or stopping the project.
This is intentional. FASO is designed to treat insufficiency as a valid result. If the prototype cannot pass repeatability, sensitivity, specificity or auditability tests, the correct outcome is redesign and rerun, not overclaiming.
£0. FASO has not raised external funding in the last 12 months. The work to date has been founder-led and self-supported, including architecture, doctrine development, prototype build, internal testing and funder/validation preparation.