Ben, your response is logical (if not correct), but the fact that many AI researchers advocate the "upbringing approach" (for other architectures) makes me very suspicious that they're anthropomorphising after all.