A nonprofit organization in New Zealand created an AI bot called Re:scam, which trolls email scammers by asking them endless questions. Here’s how it works. Once you see a suspicious email, forward it to the bot. It automatically generates responses that are programmed to include grammatical errors and humor to sound more human. If the scammer responds, the bot keeps asking the scammer more questions, annoying them with an endless conversation. The bot covers a variety of fraudulent schemes, from banking to beneficiary scams, and Western Union transfers, to romantic propositions. Here’s a demo exchange between the bot and the scammer: “Please click here for my pictures, my love,” says the scammer. “When you say nude pics do you mean .jpeg’s or .pdf’s… gifs would be ideal,” responds the bot. Billions of dollars are lost every year from email scams. The FBI estimates that between 2013 and 2016, over $1.5 billion dollars were reported lost in the US alone, due to compromised emails that target businesses and individuals who perform wire transfers.