The perfect digital fake is no longer a fake account.

It is a real person performing a tiny task inside a system they cannot see.

That is what makes the newer influence architecture harder to detect than the old bot-farm model. The post is technically authentic. The comment is written by a human. The inauthenticity lives higher up, in the instruction layer, inside closed chats and fragmented coordination. In the Moldova case described in the article, local users were recruited to carry out narrow tasks without seeing the full operation.

A useful question to ask is not, “Does this look organic?” It is: “Who benefits if hundreds of organic-looking actions converge on the same conclusion?”

fafi25.substack.com/p/w…