WHAT IS SYNTHOCRACY?
Synthocracy describes a world that already exists. Humans still vote, approve, manage, and sign. But the real work of detecting, filtering, ranking, recommending, and executing decisions increasingly runs through AI systems. The official still decides — but a model has already shaped what they see, what counts as evidence, what the default option is, and which cases ever reach them.
The word joins synthetic and -cracy (rule). It names a simple, uncomfortable shift: power can move into the machine without anyone announcing that it has.
Assisting or co-deciding?
The most important question is also the simplest. An AI tool that translates a document or fixes grammar is assisting. An AI system that ranks job applicants, scores citizens for risk, summarises evidence for a judge, or routes who gets help first is co-deciding — even if a human signs at the end. Assistance becomes co-decision the moment the system shapes what people see, trust, approve, or never review.
Synthocracy is not automatically a crime
AI can make institutions faster, fairer, and more capable. The problem isn’t that AI participates in decisions. The problem begins when that participation becomes invisible, unaccountable, or impossible to challenge. Synthocracy names a condition to be examined — not a verdict.
The questions that keep power visible
A citizen, a regulator, a journalist, or a manager can ask the same short list of any AI-mediated system: Is it assisting or co-deciding? What data does it use? Can the decision be explained? Is there a trace? Can it be appealed? Who can override it — and who can switch it off? These questions don’t require fear of AI. They require clarity.
[ See our research ] [ Read our briefings ]
