Militants in Yemen used an artificial-intelligence tool to try to develop location-guidance software for weapons, according to a threat report released by Anthropic, an incident that comes amid rising public concern that advanced AI could be harmful to human life.