... care, digital arrest threats, remote-access scams, and financial coercion. Unimind is my attempt to build an AI trust and safety infrastructure for this problem. The current direction includes TruthLayer for fact-checking text, links, and social content, and Arka, an AI call guard that detects fraud intent and coercion inside unknown calls. This is early. The goal of this launch is to open the work, find serious technical collaborators, speak with fraud/risk operators, and validate whether the strongest path ... ... verdict with links, not just a yes/no from the model. In conversations, the tricky part is that the fact is mixed with emotions and pressure (fear, urgency, threats). That s where I see TruthLayer handling the factual side, and Arka