Want to develop a risk-management framework for AI? Treat it like a human.

Here's a 7-step approach to managing AI risks using the kinds of approaches you already use for human-caused risk.