Tag archive
Model poisoning and AI supply-chain attacks sit near the top of the 2026 risk list. We set out how fine-tuning on a sealed, provenance-tracked corpus, on hardware an organisation controls, lets it vou
EU AI Act Article 10 makes data quality, provenance and sensitivity a legal obligation for high-risk systems. We examine why a cryptographically signed decision lineage answers that duty where a writt
Universities sit on some of the most valuable and most sensitive data in the world: unpublished research, patent-grade discoveries, and the personal records of tens of thousands of students. We built
(Provenance, Logistics, and the Unbroken Chain) Our last post on a Pivot Engine identified a 96 hour...
Originally published on The Searchless Journal Training data is the raw material of the AI...
I keep hearing quantum risk filed under future. It is not. My AI inference logs are being harvested today and I am required to keep them for a decade, so I make the record durable now.
If you cannot attest what went into a model, you cannot defend what comes out of it. Provenance is...
Synthetic data is now training the next generation of models. Without a chain of custody, we are...
When Andrea Bartz, Charles Graeber, and Kirk Wallace Johnson filed their class action lawsuit against...