As banks, hospitals, and governments rush to put chatbots everywhere, one question keeps getting louder: when do you have a right to talk to an actual human? From GDPRs "human intervention" safeguard to the EU AI Acts human oversight rules, the law is quietly drawing a line between helpful automation and unacceptable abandonment of people to machines.