Machines increasingly stand between what is published and the people who rely on it.
I design, build and test information structures and negotiation standards that help facts retain their context, provenance and accountability as they move through AI systems.
Who is this work for?
Factual information moves through many hands and systems. How it is represented, verified, attributed, evaluated, negotiated and acted upon depends on where you stand in its production, distribution or study.
Facts rarely remain on the page where they were published. I design schemas that attach provenance, attribution, epistemic status, evidence and correction history to the information itself—so its meaning remains available when reporting is extracted, summarised or recombined.
Published frameworks, working systems and standards efforts for keeping factual information attributable, contextual and useful as it moves between people and machines.
A metadata blueprint that gives every sentence a way to carry what kind of knowledge it holds, how it changed and how it may be reused.
A research and sparring environment for my agents and me — hosted on a Jetson Orin Nano Super and governed by a formal charter.
How should factual information be represented, tested, negotiated and acted upon in systems where humans and machines jointly produce knowledge and make decisions?
Placeholder repository used to establish the public experiment-record system before the Lab’s repositories are released.
A second placeholder showing how additional public experiments will enter the repository ledger automatically.
A third placeholder completing the repository ledger while the first public Lab experiments remain in development.
Tell me what you’re building, investigating or trying to make accountable.