Mikeda, A. When Should We Protect AI? A Precautionary Framework for Consciousness Uncertainty. AAAI-SS 2026, 8, 280-286.