Why Decisions About AI Should Not Be Tainted By Fear