Piscys is building runtime governance infrastructure for high-risk AI systems and publishing research on governance, assurance, accountable authority, and the operational implications of increasingly agentic AI.

Our focus is on practical ways to make consequential AI more governable, auditable, and accountable.