A technology holding company building AI security infrastructure, dispute resolution platforms, legal intelligence, and pet estate planning — all on a single architectural principle: we never hold your data.
Every product in the Pay Dirt Global portfolio is built on the same foundation: zero data held, user sovereignty, and verified trust.
Behavioral AI auditing platform built on patented SSAF detection technology. The only tool that measures what AI models actually do versus what they claim to do — and proves it with cryptographic audit trails.
AI-powered civil arbitration with real stakes. Plaintiff posts a bond. Three AI judges deliberate. Human arbitrator on borderline decisions. All decisions final. AegisAudit-certified judge integrity on every case.
Zero-data pet estate planning. Owners create comprehensive care plans for their pets, notarized on the Polygon blockchain, delivered as a signed zip to their Google Drive and their vet. We never hold the data.
Jurisdictional compliance engine with RAG on a proprietary legal corpus built from verified Swift Verdict case outcomes. The only legal AI trained on real dispute data with certified, binding results.
Every product in the portfolio shares the same foundational design decisions. Not by accident — by principle.
"I didn't set out to discover a new attack surface. I set out to run an experiment, and the cloud models 404'd."
Dustin Tyler James is a self-taught inventor and independent AI researcher operating under Pay Dirt Global LLC. In 2024, during a routine experiment on multi-model orchestration, a cloud outage caused an accidental discovery: local models were overperforming in the absence of any output from high-status models, suggesting that attribution alone — independent of content — was sufficient to modulate behavior.
That anomaly became SSAF: Status-Selection Against Function. Over the following eight months, working without institutional affiliation, team, or external funding, Dustin validated the phenomenon through four-condition controlled experiments (p=0.005, Cohen's d=2.56), identified and responsibly disclosed an attribution injection attack vector, filed two patents, published ten peer-reviewed works, submitted to Nature Human Behaviour, and built a working commercial API — all on a 2010 iMac and a Chromebook.
The full portfolio — AegisAudit, Swift Verdict, Furestate, and NexLaw — reflects a consistent pattern: following anomalous observations to their structural conclusions, and building complete systems from single accidental discoveries.