Independent, non-peer-reviewed research on AI safety, human agency, physical trust and enforceable authority in autonomous systems.
Each record includes a complete PDF and a transparent publication status. Repository records and permanent identifiers are linked where available.
A conditional analysis of how human interests could remain represented in machine objectives yet lose their power to constrain consequential action.
A conditional theory of how omnipresent superintelligence could displace responsibility-bearing practice and weaken the developmental foundations of human agency.
An information-theoretic and decision-theoretic framework for verification where digital representations meet decision-relevant physical states.
A conceptual hazard model, mechanism-neutral reference architecture and evaluation agenda for connecting consequential action to independently verifiable authority.