Ask it one weird question.
One crook now runs thousands of romance cons, bogus crypto tips and "wrong number" openers at once. A language model does the talking. But a clanker in a wig is still a clanker. Tell it to answer in a strict format, or switch languages mid-sentence, and it will cheerfully obey. Your actual sweetheart would not.
One weird question tells you whether you're talking to a clanker, before it talks you out of your savings or your heart. That answer is yours to keep either way.
Your reports show which tricks still beat the scammers' AI, and tie the throwaway accounts that reuse the same wallets, links and phrasing into single operations.
When a trick spreads, operators patch it. The map tracks how fast, so we always know what still works and where they're heading next.
Scammers went industrial with AI. Thousands of ten-second reports are how we fight back at the same scale.
We read the phrasing, links and wallet strings, not the name or photo, which were likely stolen from someone real.
In your own chat, at your own pace. The bot never messages anyone. There is no outbound side to it.
A couple of taps. Thousands of those and the catalog rewrites itself. You can watch it on the Living Map.